Error correction (no impact on conclusion or CPU in review): The RTX 4090 chart at 15:53 has a 3700X result that we've identified is incorrect. It is the only incorrect result. The result should read 235.9FPS AVG, 157.7FPS 1%, and 140.8FPS 0.1%. We identified this error when rerunning the CPU for other tests. We apologize for the error. Fortunately, there is no impact on the Ryzen 7000 CPUs being reviewed; however, we pride ourselves on correcting errors in the rare instances they happen so that the community has the right data. Thank you! Watch the AMD R5 7600 CPU Review: th-cam.com/video/i2XeFkhR3nA/w-d-xo.html And the R7 7700 review: th-cam.com/video/qGAwgGxJLHI/w-d-xo.html Get the LIMITED 2022 Disappointment Shirt in Cotton: store.gamersnexus.net/products/disappointment-pc-t-shirt-2022-100-cotton-black Or in a lightweight Tri-Blend! store.gamersnexus.net/products/disappointment-pc-t-shirt-2022-triblend-black
Great vid again. It's been on my mind to do an ECC enabled GPU vs non ECC ram-based GPU comparison for stability during insane rendering and 24 hour tests in the near future, unless you want to do it :)
Thanks Steve! Quick question: since the power diff between this CPU and the Intel 13700 is so great, does it make sense to add some generic annual power cost, perhaps based on say 4 hours per day at some avg kw/hr energy cost? Should electricity costs jump, this might be a handy way to get a ballpark estimate of yearly operating costs of an Eco-gaming setup vs. a full-tilt state of the art gaming powerhouse. Is the difference $100 a year? $500? More? It can also help identify parts for power conscientious builds, helpful for power limited locations i.e. off-grid solar. Helps with sizing ups power backups too. As an aside, it also helps to teach younger folks some concepts which may serve them well later in business careers. Power budgets are useful for business and engineering. Always room for another column in Excel. Keep up the great work!
Why the 7900X and 7900 even have a different name. The results here proves that the new AM5 socket doesn't make sense. They moved to AM5 for better power delivery and as we see here the no difference between the 7900 and 7900X suggests that there was no need to an improved power delivery socket
It's crazy to see the difference between CPU and GPU market nowadays. It's almost another industry... Can't believe it's the same companies working in both.
@@defnotatroll Meanwhile best mid range cpu is intel and fps king is intel :D Nvidia has zero mid to low range GPUs meanwhile AMD has. So its actually opposite what you said, AMD is facing more formidable competitor in the CPU space
Not really. Power consumption is not linear to computing power. Limiting a CPU by a small percentage will immediately indicate much lower performance. To really compare efficiency, you first need to get all CPUs up to the same compute rate. Only then can you compare... Unfortunately, the way these tests are carried out does not say much.
@@alex.germany I feel like knowing that a chip performs at 95% of the performance at 50% of the energy usage is pretty useful information to know and does in fact say a lot but okay
@@iamnotmcm No if you need somewhat exponential more power to get more out of your chip and you designed your chip to do this it will always look bad in terms of power consumption. You will either get the price for efficiency or performance not both. To get to the real capabilities of the chip there needs to be a even playingfield.
We already knew this though. These new chips are essentially just the X versions with eco mode enabled, and tests with Eco 65 show pretty much the same thing. In fact, even the 13900K matches the stock 12900K's performance with the power limit set to 85W. Zen 4 and Intel 13th gen are not inefficient at normal TDPs, only high ones.
I mean if you take 13700KF and tweak it the same way you gonna reduce consumption by about 40-50% while losing minimal performance, that's how all CPUs work but Intel doesn't care about efficiency focused chips, it let users do it, so just do your own work
@@jordanplays-transitandgame1690 Wow, a $589 can match a $429 one? So impressive. Especially when you consider that once 14th gen comes out, oopsies you can't upgrade unless you buy a whole new motherboard. Thanks, Intel!
@@GamersNexus Hi! :) Yeah, I am really eyeing the new 3D V-Cache chips for an upgrade this generation. Can't wait for your review! Keep up the great work
Man seeing the 5800x3d either being the top performer or extremely competitive in gaming scenarios is insane. With the benefit of hindsight, the 5800x3d was as close as a "future proof" a cpu you could get
Until the next X3D chips come out. I have a feeling the 3D v-cache chips are gonna be steadily generational, so it's just gonna be outpaced by iterations of itself for a while. Still, you gotta say that AMD struck gaming gold with that concept. It's damn good performance for the price. Here's hoping we see the same from the next batch. I'm worried they're gonna jack up the price though.
It's nice to see a CPU that can almost be cooled by a gentle breeze from a desk fan on the other side of the room... Especially when it packs this much compute power
If you wanted to build a production machine (say a development machine or other workstation) in a compact case, there's still not much better than this. Good price/performance, low power, very good compute power, easy to cool, just very pleasant overall.
Oh yeah, that 7900 looks nice! 😍 I like building efficient and quiet air cooled builds. I do a lot more content creation compared to gaming. Definitely looks like a good fit out of this latest round. 🤝
If you use negative curve optimiser, set a static voltage of 1.3 volts as this will give alot more tweaking room with PBO. Not many techtubers mention setting a static voltage, that's something I came across testing on my 7950X.
@@earthtaurus5515 I see, I'd have to look into that (voltage vs. frequency?) curve optimization software to understand what it is and does in practice to know what you are talking about. I normally go with things stock so it works as reliably as possibly from the start without a lot of time spent. Right now I'm on an old i7-6700 gtx 1060 build that just works super reliably though definitely showing its age in photo and video editing. I'm sure you can dial things in to make it completely stable but I'd imagine that takes some time. Thanks for the tip!
@@ScottJWaldron No problem 😁. Sure by all means you should do alot more research 👍🏽. The curve optimisation is a bios setting and you can use Ryzen Master I believe. On stock on both AM4 and AM5 way too much voltage is shoved into the cores at 1.5 volts.... 🤦🏽♂by motherboard OEMs which is absolutely nuts. Undervolting is a must on AM4 and AM5 especially if you want to avoid CPU degradation on lower quality silicon. But others will say this not an issue when it is a very much an issue. On my 7950X out of the box all cores could boost to 5.7 Ghz now after prolonged testing on stock to isolate memory issues. Now only 2 cores can boost past 5.7 Ghz. These memory issues resemble what would occur with Zen / Zen + at higher memory clock frequencies (6000+ in this case). In terms of voltage vs frequency - the traditional thinking really doesn't apply at all on AM5 and to a lesser extent AM4. As it's all about the thermal headroom, the higher it is - higher the boost on AM4. Sure you can OC AM4 as you normally would. But AM5 is a completely different beast entirely. It will hit the roof in terms of thermals if you try to OC it using the core multiplier (as you would normally do with other CPUs) - it hits 100 degrees c immediately even if the OC is set 5 or 5.3 Ghz (Therefore bclk overclocking on AM5 is quite useless for daily usage for those who do cpu intensive workloads daily). But that doesn't mean the CPU is generating heat at 100 degrees c, it's the thermal wall. Regardless it will hit that pretty quickly and start to throttle especially on anything less than a 360 AIO. On AM5 In regards to 'PBO Tuning' I recommend setting PBO boost per core (with ~25 Mhz less than limit indicated by the motherboard. As just setting a boost limit of 5.7 or 5.6 Ghz doesn't mean all cores will boost to that. For example If the motherboard reports a core can boost to 5585. Set it to 5555 at most. With Hwinfo set to 20 ms cycles you will see it will rarely boost to 5555 but it will get you over the 5.5 Ghz barrier. For better boosting performance it's best to set all cores to the same limit. Then increase the thermal limit or leave it at default of 95 degrees c. I've limited mine to 80 degrees c and max CPU package power is ~167 watts no matter the workload or duration. CCD 1 never exceeds ~87 and CCD 2 ~82 degrees. Once you've dialled in a base line then increase the boost then increase the boost of each core one by one. As at some point you will see a boost drop off. The undervolting tweaks should work on Intel's CPUs too. But I have zero experience with e-cores and settings so your mileage will vary.
Glad that the tests were consistent. I ordered the Ryzen 9 7900 (no-x) the moment it launched, using the 7600 and 7700 reviews (and LTT's 7900 review) as guidelines. I saw the pattern of 88w and equal enough performance to its X-brother. My computer is over 6 years old. The 1080 that was in it already died and I'm using a friend's 1060. Anything I bought at this point is going to be about 5x the speed of what I had when it was working. Thank you GN for these videos. They have been critical to my decision making in building my new computer.
It seems like the 7900 might be the best CPU for ultra compact SFF builds that support up to a 37mm cooler height. Can't think of any other CPU that performs this well in games while staying within thermal range of low-profile coolers.
do u think it's best to spend my money on a ryzen 7 7700/ryzen 7 7800X3D instead? I only game in 1080p but I want high framerate in my competitive 1080p games. Or should I just spend on a ryzen 5 7600X? I just want my frames to go over 300+ since Im getting a 240hz monitor.
I would have liked to see the 5950x and 5900x in the charts as well as see what the difference is in enabling PBO on the 7900X chip so we could see if the 7900X is already at max headroom in these tasks.
There are other reviews and 5950X is still ahead of it in some benchmarks but7900 runs cooler. Bottom line if you have a 5950X just keep it - doesn't worth the upgrade.
He compared them with his early 7950x, etc review. As I recall the 7900X ran blender at about 92C, compared to 95C TJmax. Don't know how I remembered details like that. So using the absolute best non-radiator sized cooler it was slightly below max temperature. Basically it's running at max.
Yeah I came to make the same comment. I thought it was very strange that they left out the 5900x and 5950x, especially on the Blender power efficiency chart where they have the R5 2600. Obviously it wouldn't make sense for most people to upgrade from a 5900x or 5950x to the 7000 series right now, but not everything needs to be hyper focused on purely purchase decisions. It's just... interesting to see how they compare. Also, even if you do want to stick to purchase decisions as a driver for the content, maybe someone who has a decent AM4 system wants to answer "should I do an in place upgrade to 5950x or jump to AM5 with the 7900?" etc...
I can remember when Athlon first came out, and I built my first PC for myself. And one of the calls to fame for AMD were their efficiency. I'm glad to see that they're going back to that.
It's interesting to see manufacturers change how boosting is intended to be used. Used to be a limited time thing just to help. Now theyre like "F IT! Have it boost to the moon and keep the chip at just under thermal throttling for maximum power"
@@recklesssquirel5962 Well Nvidia's GTX 1000 series was pretty much the first that kept boosting its frequency above just a base and a single boost frequency, kinda like how they're listed in specs, but with no extra boost on top of that.
You can power limit those as well and keep almost all of the performance... people might know this if GN talked about it, but they really are either that ignorant about efficiency or they just intentionally don't want people to know.
It's great to see a new hope for efficient parts after Intel's showing. Eco mode results for the 7900x help this, and with the rising costs of energy, its important to look out for efficiency along side cost and value. Great review GN, as always!
Bro buy a 13700KF or 12700KF and just use efficiency tweak there's plenty in TH-cam shows how, you will reduce consumption by 40-50% while losing minimal performance
@@TRX25EX I seriously need a link for that video you're talking bout. The most I can see 13700k drop without crashing is 0.055 vcore offset which results in like 30w which is quite high but a farcry from 40% let alone 50
@@quantumsage4008 Majority of videos show undervolting while getting same performance or even more, some dropped it by over 50W while getting same performance but I'm talking about losing 10% or so performance with making use of E-Cores you can find guys who done it for small ITX cases PC using small air coolers
Thanks GN! Enjoy your time off! I am excited to see the performance possibilities of the Zen 4 X3D chips in the coming months. Given the performance of the 5800X3D, the 7800X3D, 7900X3D & 7950X3D sound quite promising.
Hey there! I'm a long-time viewer of your channel, absolutely love your concise and thorough news, reviews and journalism. I have a suggestion for a new graph - idle power consumption. It's a use-case everyone ends up with one way or another, and it can vary wildly from system to system, especially with newer platforms. For home-NAS systems and even gaming machines, it would give users an idea of how much power they'll be pulling with their system when they leave it idle. I'm talking literally measuring system power draw at the windows desktop; everything stock, like-for-like components as much as possible and same GPU for every system. I think it would be a fantastic addition to your already great review process. Just a suggestion, thanks Steve.
If you are going to focus on the 7900, it would be nice to have the 5900X in ALL the comparison tables. (Blender 3.3 power efficiency, for example) Same core / thread count, and a heck of a lot cheaper if you already have AM4 compatioble gear (DDR4 RAM, motherboard)
Yeah, was kinda surprised to see a lot of 2000/3000 chips but NOT 5900 that new one directly succeeds or 5950... I know it's not worth the upgrade, but it still useful information for someone who wants to upgrade from previous gens to 5000 or 7000. I know without any benchmarks that 2600 would be destroyed in performance by 7900. What's the point of making one video for each new CPU if you don't show how it compares to last gen, especially when you cry all the time about "stagnation" and % per $ in reviews. It actually starts to irritate me a little bit with GN reviews when you have to find one specific video (usually last one) where you can see every new chip on same graph. It's fine when it's just released, but when you try to check old reviews (because not everyone buys on release) it takes 2-3 tries to see the difference and make a decision.
7900 VS 5900x 1440P Gaming 7900 is on Avg. 10-15FPS better on 4K Gaming 7900 is on Avg. 5FPS better 7900 vs 5800X3D 1080P/1440P/4K on all res. +/- 2fps Problem comes down the road...8/16 are faster but will stutter vs 12/24 same thing is in gpu VRAM 8GB is just not good for 1440p/4K ultra it stutters....
@@zdenkakoren6660 That's good data for games, I can live with 5FPS less. What about apps? As a user of Blender, it was great to see it listed in the benchmarks, but highly irritating to see the 5900x missing from the comparison tables.
The people who need a 5900x need the 16 core. The people who make their money from those upgraded already (if they didn't get threadripper for some reason). The 12 core has always been weird and kinda lacked a place in the market
I’m a little let down by the poor heat spreader design on zen 4 for compatibility, but these ‘eco’ x free CPU’s being so efficient for it to not matter is a strong case for it.
kina virus released to fk up the elections...millions dead, billion+ with heart/organ issues...evil know no bounds, 100% intentionally released...just to get one man out of office, one man they could not buy/control....Godless people will always create hell !!!
Another reason for the thick heat spreader is because the 7900x3d, etc will need extra thickness. GN ran the 7950x at 250W with excessive cooling, so only ocd overclockers will worry about that.
@@DKTD23 you don't need to enable eco mode first, the temp target is its own independant setting. Whatever limit is reached first (temp or power) will start limiting clocks then.
Thank you Tech Jesus (and team) to deliver 3 extremely high quality videos so quick! So much more useful than watching some random CES gadget no one will buy
and the fx8350 was the equivalent of intel i5 and costed half the price had a tdp of 220w and consumed 100w but disipated a lot of heat, intel i5 and i7 consumed 150-250w and still do
@@Jarandjar it didnt have 4 fake cores, it had 4 cores made out of 2 small cores,costed 200$ had the power of intel 350$ i5 or i7,people compared it with intel 1500$ 8 core and said amd is bad,then ryzen 1700 launched it had 8 cores slightly stronger than i7 8 core,i7 costed 1200$ r7 costed 200-300$
Great reviews of the non-X CPUs by the GN team. Prices are so strange for AMD CPUs right now, in some places i can get the X parts for less than the non-X parts and others there is more than $100 difference right now.
I'd have liked to see the charts consistently list the AMD and Intel equivalent prior generational chips, I don't see the 39xx or the 59xx chips show up consistently in the charts for the 7000 reviews, they are there in a few charts but I would have found it far more useful to see the differences for the previous generation 9 Series chips rather than the R5 or R7 chips which are more consistent in showing up. I sort of expected the 9 series not to be in the 7600 or 7700 reviews but had hoped that the 7900 or all R9 reviews to contain comparisons to the previous generational R9 series chips to show up in all the charts for the R9 series. This would have made the review information far more meaningful for comparison as to what all the factors are for updating generationally. I know that you put an awful lot of effort into conducting your reviews and gathering and presenting meaningful and relevant information, you do the best job in TH-cam world and I love you all for it, just some constructive criticism that's all :)
Excellent review and video. Love the detail marking the sections on the progress bar and the one on the left side of the video indicating how long till finish to show the charts. Very usefull. Thanks a lot for this fantastic work
I've been slowly upgrading my now 5 year old rig. Bought a new CPU cooler and went from 1060 3GB to 3060. Currently on a R5 2600 and this and more benchmarks like it have almost convinced me to buy a 5800X3D. Just hoping that the release of the new 7000 series X3D models will bring it's price down since 400€ is quite a stretch for a CPU.
@@PineyJustice oh, what is the task that you regularly do where 13600k uses 190w? By the way, my OC:d 5950x can also use over 220W in some very limited amount of tasks which really do not represent average day.
Thanks Steve, great content! I am now leaning, very much, towards the 7900 and am meaning to move in the next few weeks. The trouble is that I feel I had better wait until I find out about the X3D, which could blow everything out of the water.
got my 3700X for its efficiency. was holding out for a direct upscale in terms of something as efficient, and though it didn't really happen with am4, this looks nice
Shout out for letting us know you're giving your team a breather, looking forward to the "plans", but I'm more happy about the staff having some slower time after all the stuff your team pumped out in the last... geez... while... months?
I'm on a Ryzen 1 1700, and looking to upgrade to the 7900, so the performance jump should be massive, while keeping the same power efficiency. I do game a lot but also do some production, and so this one will be the best for me. But I'm interested in seeing what the 7800x3D does. Great reviews, thanks GN!
@@Apollo-Computers And we should never doubt AMD lol. Actually if they say 20% then thats probably the most it will be. I think I can do without the 7800X3d then. I do production as well so 7900 sounds good :)
@@Apollo-Computers Thanks. I think I'll just get the 7900, but I may way for a cheaper Motherboard, I don't know. I can spend all day trying to decide haha
@@mintydog06 I wanted a new system too but the motherboard prices have gotten ridiculous. I miss the days when a top end board was about half the price of a top end cpu for that board. Then we got these insane gpu prices as well. I have pretty much given up on pc's.
@@ClearAlera that may be true right now but what about in 6 months or a year when I want to go back to these reviews? Also there's definitely pros to going with the 7900. Buying an entire new system with a dead end platform is not the best advice for everyone.
@@ClearAlera Reviewers always point out that you can use newest intel chip with ddr4. But when my friend was buying a one, he found out that there does not exist any intel motherboards that support both ddr4 and ddr5. So then that makes in reality all intel cpu buyers also to go to ddr5 route as it does not make sense to buy much slower ddr4 motherboard for new cpu.
@@mikalaitio3656 Actually there are some users who still stay with the older DDR4 even now regardless of CPU manufacturer... those that use their desktop platforms for workstation duties as it's more stable to drive 4x32GB DDR4 dimms with their lower latencies, then it is to now drive 4x32GB in DDR5 in the current gen desktop platforms... though once all the early-adopters testing is done, the currently available/released DDR5 4x48GB kits running at ~7000MT/s may sway those users towards the newer gens. Heck, once/if will be getting 64GB UDIMMs running at 6000+MT/s stable those users will switch... though by that time we'll see the return of DDR5-based server-platform originating HEDT SKUs. On the other hand, half year later after R9 7900 release, the 48GB-DIMM-based kits now keep being marketed for Intel 7xx chipset platform generation based families as Zen4 now seems to be working best with approx. 6000MT/s CL30 2x32GB kits, dropping to under ~5000MT/s or failing to post with 4 high speed high capacity DIMMs. Until AMD fixes all those mem issues in Zen4 platform, those needing capacity with good performance and rock-solid stability will be choosing DDR4 generation... or/and using Intel platform.
Hey Steve and team! amazing content as always! One little thing: it is kinda hard to understand you when you say "7700x" or "7900x" because you say it too quick. I've gotten used to it over the years, but there may be people that can't follow it very well! The videos are so packed of info that you may need to rush but I just wanted to give you this feedback! Thanks for all the hard work you guys do over there!
@@Bleckyyyy even the new 3d chips will most likely run into GPU bottleneck where our "old" CPU will keep up especially at 4K. The 5800x3d will go down as a true LEGEND!
Thank you so much Computer Jesus, great info as usual. As to the sponsor, I replaced the original paste and pads on my 970gtx w/ Therm.Griz liquid metal & pads, runs like a champ!
i really look forward for more low power tests, as we can see both amd and intel push their cpus at very steep marginal gains, for example, get away with a 20 dollar low profile cooler on a 13600 or 13700 instead of using premium large tower ones, that will definitely save more as a combo.
Literally bought a new rig this week and due to receive it Monday. Started with a 7700, flip flopped over the 7700x but eventually chose the 7900. More threads, more cores, brilliant thermals and a broader range of uses which is ideal for my music production needs as well as some gaming. I have paired it with the 7800XT and 64GB of ram, hopefully this will be a great rig.
Would have been interesting to include a 7900X with target temperature set at 45 degrees (same as the 7900 got, here) for efficiency and game performance comparison.
Really fantastic review! If possible, could you also include idle and low-draw power consumption numbers? I don't often have to run any of my machines full-out for more than a few minutes at a time, so idle power consumption and efficiency matters quite a bit to my power bill.
Would like to see how the 7900 performance benchmark when underclocked to run on 45W and then running on Linux host a multiple KVM virtual machines simulating web servers or code builders for example and with 1-2 CPU's and 4-8GB of ddr5 memory,
As an itx builder for many years now, the R9-7900 (non-x) is a really good choice for a mobile gaming / workstation platform 🥰💪👍😇. If only the mobo prices were a little more down to earth 🤯🤔!
"we'll see you at the next e-sports thing" I love the dry humor guys... and the excellent graphs, great insight, the impressive methodologies.... just seriously keep up the good work!
After 7800x3d release, the price drops made things really strange at least in the UK. You can find a discounted 7900x on same price or even cheaper than non X version so I went with X version. AMD's pricing policy at launch was really unfair and remains to be so at least now for the non X versions. Thank you for the video, super helpful as always
@@masimyildiz no I used the same AIO I've been using for years which works better and quiter than a stock cooler also keeping the old PSU. Bought a very good budget B650 mobo and you don't need anything more expensive than what can be used for non-X CPUs Like I said it was cheaper than 7900. I have use cases for 7900x that justifies buying it as well and pretty much no brainer when a lesser product was more expensive. I can always underclock a higher end to lower end if watt consumption is an issue. Maybe you shouldn't try to act smart without knowing anything about the other person and embarass yourself
Sure if you limit the power draw sufficiently and handicap the CPU you can cool it by blowing air from your mouth. This is like taking a Ferrari, driving it on a 25 mph road for 2 miles a day, and saying it has affordable fuel costs.
@@FlabbyTabby LOL getting 80-90% of the performance using ONLY 43% of the power is amazing and not even close to the ridiculous comparison you are trying to make... but I get your point.
This is an excellent comparison. For the most complete picture, however, please consider adding PBO results for AMD's "X" CPUs as well as the new non-X variants.
I don't really care about energy efficiency to save power in terms of money, but the amount of lesser heat generated is an absolute game changer. Going from a 5600 to 5800x, I could def feel my room getting much hotter over time so I could only imagine how cool the 7900 would be. Great work AMD & GN team for the review.
An interesting conflating factor there is where a faster CPU enables the GPU to run quicker, the latter thus drawing more power aswell, ie. beware of attributing all of an increased heat output to the CPU when using a better SKU. Varies though depending on the game, settings, etc.
@@innocentiuslacrim2290 Likewise how can one objectively control for ambient variation, what clothes one was wearing at the time, the effect of humidity on heat sensation, any direct sunlight coming through windows, time of day, how much any entrance door was open, etc. I've been places where 52C and no humidity was much easier to tolerate than 37C and high humidity (Rhodes vs. Malta). Humidity kills me in terms of heat sensation.
Got the 7900 when I upgraded to AM5/DDR5 2 months ago and I absolutely love it. 80 minutes of music in 16/44 WAV format takes 2 minutes to convert to 320 kbps mp3, so 40x real-time! The CPU is also a very good match for my almost 3-year-old RTX 3090. Cyberpunk 2077 runs in 96 FPS average (55 fps lowest percentile) in RTX Ultra with that CPU/GPU combo.
Can’t wait to see 7950X3D vs 5950X stock for stock benchmarks - I have the latter and haven’t decided for sure, but I play Cities Skylines, which runs like Cyberpunk on AMD CPUs - best way I can describe it - or Flight Simulator. Basically, if I see the new part is 50% faster than my 5950X in at least two of the games I play, then it’s the biggest CPU jump I’ve ever seen and worth the cost of a new platform. If not, I might hold out another 12 months, like a sane person.
@@spookyskellyskeleton609 3700x change to 5800x3D, my fps +50% and I can speedup more.3D cache is very very helpful. CPU can load up to 60~70%.Fast date > more loading > more fps and speed
Yes not at full load though, you can usually see intel have lower power consumption than even amd during gaming or any other app that isn't power virus
Without the E-Cores Intel wouldn’t be able to compete with AMD in core count and multithreading. They wouldn’t be able to keep adding as many P-Cores due to die size and power.
Is there any way to set PBO at runtime or are we limited to pre-boot settings? Anxiously waiting for the reviews of the 3D versions. The wait for that EVGA season finale is also killing me slowly. The team is doing an incredible job. You were always one step above in your coverage, but are still somehow getting better.
I have been noticing that Intel chips often get recommended because of cost benefits because of the ability to use DDR4, however the tests are done with DDR5. Also the AMD 5800X3D gets touted as a contender "but on a dead end platform", where as Intel chip recommendations don't receive the same warning, even knowing Intel's propensity to change the pin count of their CPUs after 2 generations, meaning that Intel 14000 CPUs will very likely be on a new platform, with DDR 5 only.
@@mikalaitio3656 Did I say that there was? But if you are going to tell people to buy Intel chips because of the price of DDR 4, then you need to run the benchmarks on the Intel chips using DDR 4 (so a motherboard with DDR 4 slots).
There is one thing that... grind my gears watching those reviews (this is the case with many reviews of amd's non x cpus). Where the hell is 5900x/5950x datapoint. We have 3700x, 2700x, 5800X3d, 5800x - but no 5950x which is far more relevant in that comparison.
kina virus released to fk up the elections...millions dead, billion+ with heart/organ issues...evil know no bounds, 100% intentionally released...just to get one man out of office, one man they could not buy/control....Godless people will always create hell !!!
The app onto 510 E is one of my favorite processors of all time. It was a 45 white quad core laptop processor in a desktop form factor, what made it special, was that you can put it in an HTPC, NAS, or media server and it runs super quiet and super cool and has an amazing amount of power forwards TDP I was still using it up until a few months ago when the motherboard blew the USB and I haven’t found a replacement yet.
Honestly, if we are primarily comparing 7900 to 7900X, I would've liked to see 7900X PBO results in the charts as well. Because although I get the point "you can get 7900X performance by just putting 7900 on PBO", I don't think it's completely fair without including 7900X PBO as well. Not sure why that isn't included, is there something obvious I'm missing?
While this is true, scientifically speaking, GN made a video some time ago reviewing PBO on many AMD CPUs. The take-home message was: PBO almost does nothing on CPUs where thermals are not a extremely hard constraint. You may have seen the other non-X Ryzen tests, where PBO doesn't change performance (at most a few percents), basically because they were reduced from TDP 105W to 65W, the drop in thermal constraint was not even sufficient to show any difference using PBO. On the 7900 though, the performance drop from TDP 170W to 65W is big enough to notice a difference when PBO is turned on. So when the CPU is thermally unconstrained from the beginning, GN noticed that it really doesn't make any difference. That's probably why the Ryzen X-series is not tested with PBO.
@@kumarhiranya my source is brain, but it's a rare thing nowadays, I know. All X parts are only temp limited and don't reach any power constraints if there are any, that's how PBO OC is done when people mean PBO.
Great review. I really wish you would add the 13th Gen parts limited to a similar "ECO" Wattage in the BIOS so we can see how they perform in a similar setup. I know der8auer did some of this testing but would love to see it in your extensive charts
@@maxslayer11 its pretty straight forward for all BIOS'. I don't have an MSI so not sure how they handle it. But you can probably look it up on youtube for the Z690 boards from MSI.
Would love to see some of the benchmarks include stuff from a couple generations further back as context, I'm currently running a i9-7900x and would love to see how it compares to modern tests. Great work though keep it up!
Would be really interesting to see how these non x chip perform using low profile coolers. Nice to see products that doesn't push CPU cooling to the limit in sff builds.
I bought a Ryzen 9 7900 a couple of weeks ago as an upgrade from a Ryzen 7 5700G. My PC runs whisper quiet on using the very same CPU fan. And in a compact mATX system too.
Isn't it worth keeping the i7-13700k and i5-13600k for the 0.1% frame discussion alone? I mean in Far Cry 6. Going from 44 to 86, 0.1% is very significant. Why is frame consistency not talked about here. Isn't this important more than 190 vs 180 fps of the i9 vs i5
It would be interesting to include one 10 year old setup in the test charts as a reference, like the piledriver 8350 or sandybridge i7-3820.(PCIe2 and DDR3) For those of us that still aren't convinced hardware has advanced enough to economically justify an upgrade.
@@jbrone1241 Depends on the game of course, but the cost angle starts to become most amusing, eg. a XEON 1620 v3 (faster than the i7 3820) is only $8 but runs surprisingly well, likewise a 3930K costs diddly (18 UKP in the UK from CEX), while moving upwards even a used 6850K is still cheaper than a new i3 10100F. Alas in every csae it's finding a decent mbd which is difficult. MyTech, there's plenty of existing data to show the differences, but for someone who supposedly cares so much about cost that they're still using decade old hw, it would make far more sense to consider used previous gen parts anyway, especially something like a used AM4 + Zen2 combo (the 2700X often sells very cheaply for some reason, at least compared to Zen3 parts). Or for core count one can go the cheap Aliexpress/XEON route, in which case Haswell XEONs become very appealing. There are so many options if price is a prime consideration. Btw, the i7 3820 is technically SB-E, not SB, as it's for S2011 rather than S1155, though what's nice about the 3820 is it does have 40 PCIe lanes like all SB-E/IB-E CPUs, though in that regard my favourite is the unlocked 4820K (higher thermal budget per core for oc'ing, supports PCIe 3.0 native). Lastly, note that one could use an NVIDIA tool to force SB-E parts to run with PCIe 3.0, I used it with no issues. IB-E runs with 3.0 anyway, but the 4930K/4960X don't oc as easily as SB-E (there's a 5GHz Owners Club for the 3820 on OCN).
Steve when will you test idle power consumption, and gaming/ST? That's where 99% of people, especially your viewers, actually use these CPUs. Only a small fraction of people do 100% loads for extended periods.
We already do. The problem is that idle power consumption is generally so low that it amounts to noise in the measurement -- there are so many random background tasks that it's very difficult to get a true and accurate "idle" reading. Because of that, we don't present them because they are not particularly useful.
@@sowa705 But is would be useless/misleading when comparing since it's not replicable. You would be comparing apples to oranges, that at best is just bad science and at worst it can be hugely misleading and GN would be spreading "false" information overall.
I really want to see more emphasis put into the power consumption part. I know this is a "more is better" (gamers) channel, but for the stock 7900 price you get a cooler included and for a very SFF machine its unbeatable. I really want AMD to push the power limits this way for the new generation APUs, if you get amazing 6-8 core performance/watt, they can use the rest of the available power for RDNA3 iGPU cores. Imagine a SFF machine with an APU of ~95-105W (or actually more) with Graphics performance of a 3050! (fingers crossed, assuming 40-60W GPU). That is 1080p gaming without a dedicated GPU and on a
Thank you Steve, for all you do. They are not selling the non X version anymore in the netherlands which sucks. perhaps i can buy it from abroad and ship it home since even that is cheaper than getting the 7900X
Error correction (no impact on conclusion or CPU in review): The RTX 4090 chart at 15:53 has a 3700X result that we've identified is incorrect. It is the only incorrect result. The result should read 235.9FPS AVG, 157.7FPS 1%, and 140.8FPS 0.1%. We identified this error when rerunning the CPU for other tests. We apologize for the error. Fortunately, there is no impact on the Ryzen 7000 CPUs being reviewed; however, we pride ourselves on correcting errors in the rare instances they happen so that the community has the right data. Thank you!
Watch the AMD R5 7600 CPU Review: th-cam.com/video/i2XeFkhR3nA/w-d-xo.html
And the R7 7700 review: th-cam.com/video/qGAwgGxJLHI/w-d-xo.html
Get the LIMITED 2022 Disappointment Shirt in Cotton: store.gamersnexus.net/products/disappointment-pc-t-shirt-2022-100-cotton-black
Or in a lightweight Tri-Blend! store.gamersnexus.net/products/disappointment-pc-t-shirt-2022-triblend-black
Too bad apparently the leaks suggesting a 7600X3D were wrong.
Great vid again. It's been on my mind to do an ECC enabled GPU vs non ECC ram-based GPU comparison for stability during insane rendering and 24 hour tests in the near future, unless you want to do it :)
Thanks Steve! Quick question: since the power diff between this CPU and the Intel 13700 is so great, does it make sense to add some generic annual power cost, perhaps based on say 4 hours per day at some avg kw/hr energy cost?
Should electricity costs jump, this might be a handy way to get a ballpark estimate of yearly operating costs of an Eco-gaming setup vs. a full-tilt state of the art gaming powerhouse. Is the difference $100 a year? $500? More?
It can also help identify parts for power conscientious builds, helpful for power limited locations i.e. off-grid solar. Helps with sizing ups power backups too.
As an aside, it also helps to teach younger folks some concepts which may serve them well later in business careers. Power budgets are useful for business and engineering.
Always room for another column in Excel.
Keep up the great work!
Just slap Noctua on anything
Why the 7900X and 7900 even have a different name. The results here proves that the new AM5 socket doesn't make sense. They moved to AM5 for better power delivery and as we see here the no difference between the 7900 and 7900X suggests that there was no need to an improved power delivery socket
The GN team has been absolutely killing it. Seriously, you guys really have this stuff down pat.
Thank you!
@@GamersNexus I'll second that notion. Found you on the Xbox one x video. Been biblical since.
What I've known for a long time already. Welcome to the party 😃😆
But which pat?
Down Pat(rick)
It's crazy to see the difference between CPU and GPU market nowadays. It's almost another industry... Can't believe it's the same companies working in both.
Nvidia are better at making great GPUs than Intel are at making great CPUs. So AMD are up against a more formidable competitor in the GPU space
@@defnotatroll Meanwhile best mid range cpu is intel and fps king is intel :D Nvidia has zero mid to low range GPUs meanwhile AMD has.
So its actually opposite what you said, AMD is facing more formidable competitor in the CPU space
@@DuBstep115 Mid range?
Almost high end. So much talk about the i5-13600K but the i7-12700K exists.
@@saricubra2867 Yes, mid range aka 3060-3060ti. If you have any brains you buy 6700xt for like $50-100 cheaper
@@DuBstep115 if you have brains you get a second hand card
200w less than a 13700k and 95% if the performance. Impressive.
now THAT is some crazy efficiency indeed!
Not really. Power consumption is not linear to computing power. Limiting a CPU by a small percentage will immediately indicate much lower performance. To really compare efficiency, you first need to get all CPUs up to the same compute rate. Only then can you compare... Unfortunately, the way these tests are carried out does not say much.
@@alex.germany I feel like knowing that a chip performs at 95% of the performance at 50% of the energy usage is pretty useful information to know and does in fact say a lot but okay
@@iamnotmcm No if you need somewhat exponential more power to get more out of your chip and you designed your chip to do this it will always look bad in terms of power consumption. You will either get the price for efficiency or performance not both. To get to the real capabilities of the chip there needs to be a even playingfield.
@@sigmundfreud2443 so? that's useful information to people. the playing field is power consumption.
6:55 An 88w 7900 matching the performance of a _stock_ 12900k is impressive.
We already knew this though. These new chips are essentially just the X versions with eco mode enabled, and tests with Eco 65 show pretty much the same thing. In fact, even the 13900K matches the stock 12900K's performance with the power limit set to 85W. Zen 4 and Intel 13th gen are not inefficient at normal TDPs, only high ones.
I mean if you take 13700KF and tweak it the same way you gonna reduce consumption by about 40-50% while losing minimal performance, that's how all CPUs work but Intel doesn't care about efficiency focused chips, it let users do it, so just do your own work
Well duh that's multithreaded a 13900k can do it with 65w
@@jordanplays-transitandgame1690 Wow, a $589 can match a $429 one? So impressive. Especially when you consider that once 14th gen comes out, oopsies you can't upgrade unless you buy a whole new motherboard. Thanks, Intel!
@@AlexanderPavel 7900 and 7900x are 400$ 13900k is 600+$
Nice, exciting to see efficiency gains like this! Wonder how the 7900X3D is going to perform...
Really looking forward to testing the new X3D parts! Won't be long now.
I expect the 7900X3D to excel at emptying my bank account.
@@GamersNexus Hi! :) Yeah, I am really eyeing the new 3D V-Cache chips for an upgrade this generation. Can't wait for your review! Keep up the great work
@@GamersNexus X3D have performance improvements for productivity work or is it solely gaming improvements?
@@ramicollo only improvments in tasks that actually use the cache...
The ENTIRE team is doing great, hope Y'all good. 👍
Man seeing the 5800x3d either being the top performer or extremely competitive in gaming scenarios is insane. With the benefit of hindsight, the 5800x3d was as close as a "future proof" a cpu you could get
It's not even a year old, wtf are you smoking?
Reminiscent of the i7 2600k
@@SeanCMonahan Yeah, same thing I was thinking. The 2600K lived on for ages in gaming. Really happy with getting a 5800X3D.
@Sean Monahan pretty much until the 8700k/3600 it was plenty.
Until the next X3D chips come out. I have a feeling the 3D v-cache chips are gonna be steadily generational, so it's just gonna be outpaced by iterations of itself for a while. Still, you gotta say that AMD struck gaming gold with that concept. It's damn good performance for the price. Here's hoping we see the same from the next batch. I'm worried they're gonna jack up the price though.
It's nice to see a CPU that can almost be cooled by a gentle breeze from a desk fan on the other side of the room... Especially when it packs this much compute power
If you wanted to build a production machine (say a development machine or other workstation) in a compact case, there's still not much better than this. Good price/performance, low power, very good compute power, easy to cool, just very pleasant overall.
Not really. I still did a Titan AIO. It's cooler and quieter than the included air wraith cooler by about 5C. Probably even more once it gets cooking.
The most impressive thing is you can easily run 7900 with included wraith prism cooler
No, the most impressive thing is that you can see GN team in the next e-sports thing.
@@sirius4k esport* not e-sports. Yikes.
@@MarioWizzz That's what I said, e-sports.
@@MarioWizzz Depends on what he means, e-sport = athlete or e-sport = a good sport on youtube.
at 50 degrees C, you can run it with a table fan blowing on it.
Oh yeah, that 7900 looks nice! 😍 I like building efficient and quiet air cooled builds. I do a lot more content creation compared to gaming. Definitely looks like a good fit out of this latest round. 🤝
I love AMD for making these cheap CPUs that sit out there and do their job while staying cool, quiet and power efficient.
If you use negative curve optimiser, set a static voltage of 1.3 volts as this will give alot more tweaking room with PBO. Not many techtubers mention setting a static voltage, that's something I came across testing on my 7950X.
@@earthtaurus5515 I see, I'd have to look into that (voltage vs. frequency?) curve optimization software to understand what it is and does in practice to know what you are talking about. I normally go with things stock so it works as reliably as possibly from the start without a lot of time spent. Right now I'm on an old i7-6700 gtx 1060 build that just works super reliably though definitely showing its age in photo and video editing. I'm sure you can dial things in to make it completely stable but I'd imagine that takes some time. Thanks for the tip!
@@ScottJWaldron No problem 😁. Sure by all means you should do alot more research 👍🏽. The curve optimisation is a bios setting and you can use Ryzen Master I believe.
On stock on both AM4 and AM5 way too much voltage is shoved into the cores at 1.5 volts.... 🤦🏽♂by motherboard OEMs which is absolutely nuts. Undervolting is a must on AM4 and AM5 especially if you want to avoid CPU degradation on lower quality silicon. But others will say this not an issue when it is a very much an issue.
On my 7950X out of the box all cores could boost to 5.7 Ghz now after prolonged testing on stock to isolate memory issues. Now only 2 cores can boost past 5.7 Ghz. These memory issues resemble what would occur with Zen / Zen + at higher memory clock frequencies (6000+ in this case).
In terms of voltage vs frequency - the traditional thinking really doesn't apply at all on AM5 and to a lesser extent AM4. As it's all about the thermal headroom, the higher it is - higher the boost on AM4. Sure you can OC AM4 as you normally would.
But AM5 is a completely different beast entirely. It will hit the roof in terms of thermals if you try to OC it using the core multiplier (as you would normally do with other CPUs) - it hits 100 degrees c immediately even if the OC is set 5 or 5.3 Ghz (Therefore bclk overclocking on AM5 is quite useless for daily usage for those who do cpu intensive workloads daily). But that doesn't mean the CPU is generating heat at 100 degrees c, it's the thermal wall.
Regardless it will hit that pretty quickly and start to throttle especially on anything less than a 360 AIO.
On AM5 In regards to 'PBO Tuning' I recommend setting PBO boost per core (with ~25 Mhz less than limit indicated by the motherboard. As just setting a boost limit of 5.7 or 5.6 Ghz doesn't mean all cores will boost to that.
For example If the motherboard reports a core can boost to 5585. Set it to 5555 at most. With Hwinfo set to 20 ms cycles you will see it will rarely boost to 5555 but it will get you over the 5.5 Ghz barrier.
For better boosting performance it's best to set all cores to the same limit. Then increase the thermal limit or leave it at default of 95 degrees c. I've limited mine to 80 degrees c and max CPU package power is ~167 watts no matter the workload or duration. CCD 1 never exceeds ~87 and CCD 2 ~82 degrees.
Once you've dialled in a base line then increase the boost then increase the boost of each core one by one. As at some point you will see a boost drop off.
The undervolting tweaks should work on Intel's CPUs too. But I have zero experience with e-cores and settings so your mileage will vary.
Glad that the tests were consistent. I ordered the Ryzen 9 7900 (no-x) the moment it launched, using the 7600 and 7700 reviews (and LTT's 7900 review) as guidelines. I saw the pattern of 88w and equal enough performance to its X-brother. My computer is over 6 years old. The 1080 that was in it already died and I'm using a friend's 1060. Anything I bought at this point is going to be about 5x the speed of what I had when it was working. Thank you GN for these videos. They have been critical to my decision making in building my new computer.
It seems like the 7900 might be the best CPU for ultra compact SFF builds that support up to a 37mm cooler height. Can't think of any other CPU that performs this well in games while staying within thermal range of low-profile coolers.
do u think it's best to spend my money on a ryzen 7 7700/ryzen 7 7800X3D instead? I only game in 1080p but I want high framerate in my competitive 1080p games. Or should I just spend on a ryzen 5 7600X? I just want my frames to go over 300+ since Im getting a 240hz monitor.
@@sobangcha3042 7800x3d.if.you r doing gaming at 1080p
I would have liked to see the 5950x and 5900x in the charts as well as see what the difference is in enabling PBO on the 7900X chip so we could see if the 7900X is already at max headroom in these tasks.
Yup. A single generation jump is not worth the upgrade with the high prices of AM5 motherboards and all the DDR5 sh!t.
@@HanSolo__ Not worth it *yet*. In a year or so, might be a good upgrade.
There are other reviews and 5950X is still ahead of it in some benchmarks but7900 runs cooler. Bottom line if you have a 5950X just keep it - doesn't worth the upgrade.
He compared them with his early 7950x, etc review. As I recall the 7900X ran blender at about 92C, compared to 95C TJmax. Don't know how I remembered details like that. So using the absolute best non-radiator sized cooler it was slightly below max temperature. Basically it's running at max.
Yeah I came to make the same comment. I thought it was very strange that they left out the 5900x and 5950x, especially on the Blender power efficiency chart where they have the R5 2600. Obviously it wouldn't make sense for most people to upgrade from a 5900x or 5950x to the 7000 series right now, but not everything needs to be hyper focused on purely purchase decisions. It's just... interesting to see how they compare. Also, even if you do want to stick to purchase decisions as a driver for the content, maybe someone who has a decent AM4 system wants to answer "should I do an in place upgrade to 5950x or jump to AM5 with the 7900?" etc...
I can remember when Athlon first came out, and I built my first PC for myself. And one of the calls to fame for AMD were their efficiency. I'm glad to see that they're going back to that.
I really appreciated the R7 3700x cpu being included on the charts. Reminds me that my bottleneck for gaming is currently my 1060, not the 3700x.
@@dateinamwahlang6463my goodness hahaha. Atleast you wont get a cpu bottleneck anytime soon lol.
I'm glad they included the R7 2700 results lol, as I will be upgrading from that. Looks to be around double to triple the CPU performance.
It's interesting to see manufacturers change how boosting is intended to be used. Used to be a limited time thing just to help. Now theyre like "F IT! Have it boost to the moon and keep the chip at just under thermal throttling for maximum power"
Essentially what GPUs did years ago too...
@@shawnpitman876 i haven't been in the pc space for very long
@@recklesssquirel5962 Well Nvidia's GTX 1000 series was pretty much the first that kept boosting its frequency above just a base and a single boost frequency, kinda like how they're listed in specs, but with no extra boost on top of that.
You can power limit those as well and keep almost all of the performance... people might know this if GN talked about it, but they really are either that ignorant about efficiency or they just intentionally don't want people to know.
@@Lambda.Function Because they don’t care as they are trying to sell these products as an “influencer”…
Would have loved to see 5900X and 5950X on the charts for comparison but a great vid all the same.
The expression on Steve's face in the thumbnail is all I usually need to know the conclusion of the video.
It's great to see a new hope for efficient parts after Intel's showing. Eco mode results for the 7900x help this, and with the rising costs of energy, its important to look out for efficiency along side cost and value. Great review GN, as always!
rising cost of energy ? Yea but for Europeans 🤣
@@kipa_chu energy bill last 3 months has been ~27 euro each month lol
Bro buy a 13700KF or 12700KF and just use efficiency tweak there's plenty in TH-cam shows how, you will reduce consumption by 40-50% while losing minimal performance
@@TRX25EX I seriously need a link for that video you're talking bout. The most I can see 13700k drop without crashing is 0.055 vcore offset which results in like 30w which is quite high but a farcry from 40% let alone 50
@@quantumsage4008 Majority of videos show undervolting while getting same performance or even more, some dropped it by over 50W while getting same performance but I'm talking about losing 10% or so performance with making use of E-Cores you can find guys who done it for small ITX cases PC using small air coolers
Thanks GN! Enjoy your time off!
I am excited to see the performance possibilities of the Zen 4 X3D chips in the coming months. Given the performance of the 5800X3D, the 7800X3D, 7900X3D & 7950X3D sound quite promising.
Hey there! I'm a long-time viewer of your channel, absolutely love your concise and thorough news, reviews and journalism.
I have a suggestion for a new graph - idle power consumption. It's a use-case everyone ends up with one way or another, and it can vary wildly from system to system, especially with newer platforms. For home-NAS systems and even gaming machines, it would give users an idea of how much power they'll be pulling with their system when they leave it idle.
I'm talking literally measuring system power draw at the windows desktop; everything stock, like-for-like components as much as possible and same GPU for every system.
I think it would be a fantastic addition to your already great review process. Just a suggestion, thanks Steve.
If you are going to focus on the 7900, it would be nice to have the 5900X in ALL the comparison tables. (Blender 3.3 power efficiency, for example)
Same core / thread count, and a heck of a lot cheaper if you already have AM4 compatioble gear (DDR4 RAM, motherboard)
Likewise good to include the 5950X which was a much better binned part than the 5900X, though alas pricing for both has become unpleasant.
Yeah, was kinda surprised to see a lot of 2000/3000 chips but NOT 5900 that new one directly succeeds or 5950... I know it's not worth the upgrade, but it still useful information for someone who wants to upgrade from previous gens to 5000 or 7000. I know without any benchmarks that 2600 would be destroyed in performance by 7900. What's the point of making one video for each new CPU if you don't show how it compares to last gen, especially when you cry all the time about "stagnation" and % per $ in reviews. It actually starts to irritate me a little bit with GN reviews when you have to find one specific video (usually last one) where you can see every new chip on same graph. It's fine when it's just released, but when you try to check old reviews (because not everyone buys on release) it takes 2-3 tries to see the difference and make a decision.
7900 VS 5900x
1440P Gaming 7900 is on Avg. 10-15FPS better on 4K Gaming 7900 is on Avg. 5FPS better
7900 vs 5800X3D 1080P/1440P/4K on all res. +/- 2fps
Problem comes down the road...8/16 are faster but will stutter vs 12/24 same thing is in gpu VRAM 8GB is just not good for 1440p/4K ultra it stutters....
@@zdenkakoren6660 That's good data for games, I can live with 5FPS less. What about apps?
As a user of Blender, it was great to see it listed in the benchmarks, but highly irritating to see the 5900x missing from the comparison tables.
The people who need a 5900x need the 16 core. The people who make their money from those upgraded already (if they didn't get threadripper for some reason). The 12 core has always been weird and kinda lacked a place in the market
I’m a little let down by the poor heat spreader design on zen 4 for compatibility, but these ‘eco’ x free CPU’s being so efficient for it to not matter is a strong case for it.
kina virus released to fk up the elections...millions dead, billion+ with heart/organ issues...evil know no bounds, 100% intentionally released...just to get one man out of office, one man they could not buy/control....Godless people will always create hell !!!
the x ones you can choose in bios temp 70 and it never goes over and still boosts to 5.5ghz
@@Noooo23523 just temp setting or do you need to enable eco mode first
Another reason for the thick heat spreader is because the 7900x3d, etc will need extra thickness. GN ran the 7950x at 250W with excessive cooling, so only ocd overclockers will worry about that.
@@DKTD23 you don't need to enable eco mode first, the temp target is its own independant setting. Whatever limit is reached first (temp or power) will start limiting clocks then.
Thank you Tech Jesus (and team) to deliver 3 extremely high quality videos so quick! So much more useful than watching some random CES gadget no one will buy
Are you on reddit? Because they refer to Steve as tech Jesus over there as well.
@@drg19841 I believe it started here, but not sure
This is the same company that put out the FX-9590. They've made tremendous strides.
and the fx8350 was the equivalent of intel i5 and costed half the price had a tdp of 220w and consumed 100w but disipated a lot of heat, intel i5 and i7 consumed 150-250w and still do
Meanwhile Intel, the company that made Netburst (Pentium 4). Still waiting for them to hit 10 GHz.
@@Radovanslav fx overclocked to 9ghz 10 years ago
@@Noooo23523 Yeah but it also had 4 fake cores 😂
@@Jarandjar it didnt have 4 fake cores, it had 4 cores made out of 2 small cores,costed 200$ had the power of intel 350$ i5 or i7,people compared it with intel 1500$ 8 core and said amd is bad,then ryzen 1700 launched it had 8 cores slightly stronger than i7 8 core,i7 costed 1200$ r7 costed 200-300$
This is the one im getting glad I waited . Great job as always Steve and staff !
Great reviews of the non-X CPUs by the GN team.
Prices are so strange for AMD CPUs right now, in some places i can get the X parts for less than the non-X parts and others there is more than $100 difference right now.
Yea I'm waiting for the review of 7900, super power with much lower temps, very attractive for SFF build. thanks for your effort!
Hope that the 7000X3D parts show some real "worthy" upgrades over everything else!
Well if the 5800X3D is anything to go by it should.
@Just for LOLZ and that seems to be fixed this time around going off the info we have right now.
@@republicdecease Only kinda fixed. You have the full boost only in the chiplet w/o the extra cache.
This series of reviews has been insanely interesting. I've learned a lot.
Great job as always! Would love to see the 5900X compared but good job all the same
Thanks for all the great coverage Steve and crew at GN!
I'd have liked to see the charts consistently list the AMD and Intel equivalent prior generational chips, I don't see the 39xx or the 59xx chips show up consistently in the charts for the 7000 reviews, they are there in a few charts but I would have found it far more useful to see the differences for the previous generation 9 Series chips rather than the R5 or R7 chips which are more consistent in showing up. I sort of expected the 9 series not to be in the 7600 or 7700 reviews but had hoped that the 7900 or all R9 reviews to contain comparisons to the previous generational R9 series chips to show up in all the charts for the R9 series. This would have made the review information far more meaningful for comparison as to what all the factors are for updating generationally. I know that you put an awful lot of effort into conducting your reviews and gathering and presenting meaningful and relevant information, you do the best job in TH-cam world and I love you all for it, just some constructive criticism that's all :)
The Guru3D review does have some data of such kind.
This is the first time I've seen a video shot on this set. It looks great!
Interesting CPU selection on the charts, would have expected to see some more last gen options of similar tier (5900x/5950x)
they don't have enough people to retest everything with the 4090 and/or the latest version of apps/games for all the cpus they have unfortunately
@@damasterpiece08 3090 was used. Charts in video.
Excellent review and video. Love the detail marking the sections on the progress bar and the one on the left side of the video indicating how long till finish to show the charts. Very usefull. Thanks a lot for this fantastic work
I've been slowly upgrading my now 5 year old rig. Bought a new CPU cooler and went from 1060 3GB to 3060. Currently on a R5 2600 and this and more benchmarks like it have almost convinced me to buy a 5800X3D. Just hoping that the release of the new 7000 series X3D models will bring it's price down since 400€ is quite a stretch for a CPU.
7700x better than 13600k at same price, rx6600xt costs half than rtx3060 and is slightly stronger,rtx3060 costs like rx6800
@@username8644 am5 is 30-100$ more than am4,you are paying for 5 m.2 pcie 5.0 2.5gb inthernet and top quality sound board on all models
@@Noooo23523 7700x is not better than 13600k :-)
@@innocentiuslacrim2290 Sure it is, but only if you're pretending that the i5 doesn't draw 190 watts.
@@PineyJustice oh, what is the task that you regularly do where 13600k uses 190w? By the way, my OC:d 5950x can also use over 220W in some very limited amount of tasks which really do not represent average day.
Thanks Steve, great content!
I am now leaning, very much, towards the 7900 and am meaning to move in the next few weeks. The trouble is that I feel I had better wait until I find out about the X3D, which could blow everything out of the water.
got my 3700X for its efficiency. was holding out for a direct upscale in terms of something as efficient, and though it didn't really happen with am4, this looks nice
Same here. I wasn't happy to see the big jump in TDPs this gen but these non-X variants change everything.
Thanks Steve - just the data I needed to decide on the CPU for my next system build - perfect!
I'm liking that set with the wall of monitors behind you. :)
Shout out for letting us know you're giving your team a breather, looking forward to the "plans", but I'm more happy about the staff having some slower time after all the stuff your team pumped out in the last... geez... while... months?
Are you sure you’re not talking about TDP? It’s AMD, they’re not even sure! Thanks, Steve!
I'm on a Ryzen 1 1700, and looking to upgrade to the 7900, so the performance jump should be massive, while keeping the same power efficiency. I do game a lot but also do some production, and so this one will be the best for me. But I'm interested in seeing what the 7800x3D does.
Great reviews, thanks GN!
I believe AMD claimed 7800x3d is about 20% faster than the 5800x3d. But of course that's what AMD says...
@@Apollo-Computers And we should never doubt AMD lol. Actually if they say 20% then thats probably the most it will be.
I think I can do without the 7800X3d then. I do production as well so 7900 sounds good :)
@@mintydog06 they only showed like 4 games and the highest was like 23%. But yea just gonna have to wait for real benchmarks.
@@Apollo-Computers Thanks. I think I'll just get the 7900, but I may way for a cheaper Motherboard, I don't know. I can spend all day trying to decide haha
@@mintydog06 I wanted a new system too but the motherboard prices have gotten ridiculous. I miss the days when a top end board was about half the price of a top end cpu for that board. Then we got these insane gpu prices as well. I have pretty much given up on pc's.
The 5900X is the main processor I would want to compare to the 7900. I kept looking for it thinking I was missing it in the charts at first.
Given the costs of AM5 and DDR5 the performance uplift of a 7900 over a 5900x is almost certainly not worth it in either gaming or productivity.
@@ClearAlera that may be true right now but what about in 6 months or a year when I want to go back to these reviews? Also there's definitely pros to going with the 7900. Buying an entire new system with a dead end platform is not the best advice for everyone.
@@cosmic_drew just sell your pc and you don’t have to worry
@@ClearAlera Reviewers always point out that you can use newest intel chip with ddr4. But when my friend was buying a one, he found out that there does not exist any intel motherboards that support both ddr4 and ddr5. So then that makes in reality all intel cpu buyers also to go to ddr5 route as it does not make sense to buy much slower ddr4 motherboard for new cpu.
@@mikalaitio3656 Actually there are some users who still stay with the older DDR4 even now regardless of CPU manufacturer... those that use their desktop platforms for workstation duties as it's more stable to drive 4x32GB DDR4 dimms with their lower latencies, then it is to now drive 4x32GB in DDR5 in the current gen desktop platforms... though once all the early-adopters testing is done, the currently available/released DDR5 4x48GB kits running at ~7000MT/s may sway those users towards the newer gens. Heck, once/if will be getting 64GB UDIMMs running at 6000+MT/s stable those users will switch... though by that time we'll see the return of DDR5-based server-platform originating HEDT SKUs. On the other hand, half year later after R9 7900 release, the 48GB-DIMM-based kits now keep being marketed for Intel 7xx chipset platform generation based families as Zen4 now seems to be working best with approx. 6000MT/s CL30 2x32GB kits, dropping to under ~5000MT/s or failing to post with 4 high speed high capacity DIMMs. Until AMD fixes all those mem issues in Zen4 platform, those needing capacity with good performance and rock-solid stability will be choosing DDR4 generation... or/and using Intel platform.
Great info, especially the table about efficiency of performing the same work. Thanks bro!
Hey Steve and team! amazing content as always! One little thing: it is kinda hard to understand you when you say "7700x" or "7900x" because you say it too quick. I've gotten used to it over the years, but there may be people that can't follow it very well! The videos are so packed of info that you may need to rush but I just wanted to give you this feedback! Thanks for all the hard work you guys do over there!
I'm glad I got my 7900X on sale. It was $449 on Amazon for the holidays. thats the MSRP for the 7900, so I'm happy about it.
As an gamer only and owner of a 5800x3d I'm loving those gaming benchmarks :-)
It's a beast
I'm also smiling as it beats 13900K in many games :D PS: also a proud owner of 5800x3d :D
@@Bleckyyyy even the new 3d chips will most likely run into GPU bottleneck where our "old" CPU will keep up especially at 4K. The 5800x3d will go down as a true LEGEND!
Especially in VR sim games, is the best cpu at the moment, by a long margin.
Thank you so much Computer Jesus, great info as usual. As to the sponsor, I replaced the original paste and pads on my 970gtx w/ Therm.Griz liquid metal & pads, runs like a champ!
i really look forward for more low power tests, as we can see both amd and intel push their cpus at very steep marginal gains, for example, get away with a 20 dollar low profile cooler on a 13600 or 13700 instead of using premium large tower ones, that will definitely save more as a combo.
Literally bought a new rig this week and due to receive it Monday. Started with a 7700, flip flopped over the 7700x but eventually chose the 7900. More threads, more cores, brilliant thermals and a broader range of uses which is ideal for my music production needs as well as some gaming. I have paired it with the 7800XT and 64GB of ram, hopefully this will be a great rig.
Congrats
Would have been interesting to include a 7900X with target temperature set at 45 degrees (same as the 7900 got, here) for efficiency and game performance comparison.
"... sometimes it rendering more crash reports than frames" -- ouch, that hits home real hard!
Really fantastic review! If possible, could you also include idle and low-draw power consumption numbers? I don't often have to run any of my machines full-out for more than a few minutes at a time, so idle power consumption and efficiency matters quite a bit to my power bill.
When I saw the 88W vs 230W on the 7900 it was a no brainer for me! I'm glad to see a full review!
Would like to see how the 7900 performance benchmark when underclocked to run on 45W and then running on Linux host a multiple KVM virtual machines simulating web servers or code builders for example and with 1-2 CPU's and 4-8GB of ddr5 memory,
But did they do anything about the AM5 MASSIVE power draw ( waste) at or near idle?
@@tilapiadave3234 haven't heard about that I'll do some research! Thanks for keeping me in the loop.
As an itx builder for many years now, the R9-7900 (non-x) is a really good choice for a mobile gaming / workstation platform 🥰💪👍😇. If only the mobo prices were a little more down to earth 🤯🤔!
The good thing is when the cheap and nasty motherboards arrive this should work on them fine because the power is so low
Hello, do you think that this processor is not suitable for demanding games? Need your opinion:)
I love your recent work, great job! I do want to mention that the monitors in the background of the main set is REALLY distracting for me.
It would be good to have a video about the new 7000 CPUs with their stock coolers, to see how far PBO can go in that case.
It won't go far, because the stock coolers are crap. They sure don't cool 200W.
"we'll see you at the next e-sports thing"
I love the dry humor guys... and the excellent graphs, great insight, the impressive methodologies.... just seriously keep up the good work!
i wish the 5900x was represented in more of the graphs, but interesting part.
Benchmark curves (score vs power) would have been good to include. As always, great work showing so much detail!
After 7800x3d release, the price drops made things really strange at least in the UK. You can find a discounted 7900x on same price or even cheaper than non X version so I went with X version. AMD's pricing policy at launch was really unfair and remains to be so at least now for the non X versions. Thank you for the video, super helpful as always
And you bought a cooler, a mid/high end motherboard, and you need nearly 100 watt more power on psu. 🤦🏻♂
@@masimyildiz no I used the same AIO I've been using for years which works better and quiter than a stock cooler also keeping the old PSU.
Bought a very good budget B650 mobo and you don't need anything more expensive than what can be used for non-X CPUs
Like I said it was cheaper than 7900.
I have use cases for 7900x that justifies buying it as well and pretty much no brainer when a lesser product was more expensive.
I can always underclock a higher end to lower end if watt consumption is an issue.
Maybe you shouldn't try to act smart without knowing anything about the other person and embarass yourself
The cup coaster looks amazing with the processor sitting on top of it. Almost looks like it's made for it!
I'd love to see if a wraith stealth cooler could keep the 7900 under 80C. Could be an amazing CPU for tiny air cooled builds.
Sure if you limit the power draw sufficiently and handicap the CPU you can cool it by blowing air from your mouth.
This is like taking a Ferrari, driving it on a 25 mph road for 2 miles a day, and saying it has affordable fuel costs.
@@FlabbyTabby LOL getting 80-90% of the performance using ONLY 43% of the power is amazing and not even close to the ridiculous comparison you are trying to make... but I get your point.
This is an excellent comparison. For the most complete picture, however, please consider adding PBO results for AMD's "X" CPUs as well as the new non-X variants.
I don't really care about energy efficiency to save power in terms of money, but the amount of lesser heat generated is an absolute game changer. Going from a 5600 to 5800x, I could def feel my room getting much hotter over time so I could only imagine how cool the 7900 would be. Great work AMD & GN team for the review.
An interesting conflating factor there is where a faster CPU enables the GPU to run quicker, the latter thus drawing more power aswell, ie. beware of attributing all of an increased heat output to the CPU when using a better SKU. Varies though depending on the game, settings, etc.
How could you feel your room getting "much hotter" with a few tens of watts difference during gaming and even less during most other tasks?
@@innocentiuslacrim2290 Likewise how can one objectively control for ambient variation, what clothes one was wearing at the time, the effect of humidity on heat sensation, any direct sunlight coming through windows, time of day, how much any entrance door was open, etc.
I've been places where 52C and no humidity was much easier to tolerate than 37C and high humidity (Rhodes vs. Malta). Humidity kills me in terms of heat sensation.
@@mapesdhs597 For sure, that's a great point. I had the same GPU (3080) so that def could have been it.
9:05 - That caught me off guard. Excellent writing!
I think it would also be cool to do a non-X pbo vs an X pbo comparison, to see if it makes an sense at all to buy an X cpu
Got the 7900 when I upgraded to AM5/DDR5 2 months ago and I absolutely love it. 80 minutes of music in 16/44 WAV format takes 2 minutes to convert to 320 kbps mp3, so 40x real-time! The CPU is also a very good match for my almost 3-year-old RTX 3090. Cyberpunk 2077 runs in 96 FPS average (55 fps lowest percentile) in RTX Ultra with that CPU/GPU combo.
whats your cpu temp when gaming?
Can’t wait to see 7950X3D vs 5950X stock for stock benchmarks - I have the latter and haven’t decided for sure, but I play Cities Skylines, which runs like Cyberpunk on AMD CPUs - best way I can describe it - or Flight Simulator. Basically, if I see the new part is 50% faster than my 5950X in at least two of the games I play, then it’s the biggest CPU jump I’ve ever seen and worth the cost of a new platform. If not, I might hold out another 12 months, like a sane person.
I do not see my cpu (5950x) being the limiting factor of my Cyberpunk gaming. If I lower graphical settings or resolution, I get more fps :-)
Skylines should not be an issue for your cpu at all
@@spookyskellyskeleton609 3700x change to 5800x3D, my fps +50% and I can speedup more.3D cache is very very helpful. CPU can load up to 60~70%.Fast date > more loading > more fps and speed
@@DGCNYO 5950x is lot faster than 3700x, you really can't compare them. Its a flagship processor for 5000 series
@@spookyskellyskeleton609 I said 5800x3D....In cities skyline 5950x not fast, because cache issue
I'm always amazed at the quality of your reviews, keep up the good work
Good to see those E cores really helping with efficiency! 🤣
Yes not at full load though, you can usually see intel have lower power consumption than even amd during gaming or any other app that isn't power virus
E cores are in reality just for making them better at production tasks right?
Without the E-Cores Intel wouldn’t be able to compete with AMD in core count and multithreading. They wouldn’t be able to keep adding as many P-Cores due to die size and power.
@@w04h even in gaming it draws more, but the diffrence is way less
@@w04h not really you are probably thinking about 12th gen which consumed less at idle but 13th gen just chugs more for more performance
Thanks for the review, Steve & gang!
Is there any way to set PBO at runtime or are we limited to pre-boot settings?
Anxiously waiting for the reviews of the 3D versions. The wait for that EVGA season finale is also killing me slowly.
The team is doing an incredible job. You were always one step above in your coverage, but are still somehow getting better.
PBO settings can be altered at runtime with appropriate software.
AMD provides Ryzen Master for Windows, which allows changing OC settings at runtime
@@niter43 Or just use PBO Tuner, and task scheduler to set settings at windows startup.
Another great analysis. The only thing missing is a powerful rotating orb on one of the monitors in the background array.
At least the CPU side of things is in a healthier state than the GPU side. Wish those AM5 motherboards would drop in price however.
I was just looking at doing a M-ITX build and power consumption/heat is important. This is just what I needed.
I have been noticing that Intel chips often get recommended because of cost benefits because of the ability to use DDR4, however the tests are done with DDR5. Also the AMD 5800X3D gets touted as a contender "but on a dead end platform", where as Intel chip recommendations don't receive the same warning, even knowing Intel's propensity to change the pin count of their CPUs after 2 generations, meaning that Intel 14000 CPUs will very likely be on a new platform, with DDR 5 only.
Also there is no intel motherboards that support both the ddr4 or ddr5.
@@mikalaitio3656 Did I say that there was? But if you are going to tell people to buy Intel chips because of the price of DDR 4, then you need to run the benchmarks on the Intel chips using DDR 4 (so a motherboard with DDR 4 slots).
Honestly, i would love to see CPUs with this kind of power efficiency + 3D V-cache from AMD
There is one thing that... grind my gears watching those reviews (this is the case with many reviews of amd's non x cpus). Where the hell is 5900x/5950x datapoint. We have 3700x, 2700x, 5800X3d, 5800x - but no 5950x which is far more relevant in that comparison.
No 3950X either!
This channel is a service. Thank you! I await my fancy desk mat.
I hope you guys never change. You're the only PC parts reviewer that I trust 100% and I hope that never changes.
Thanks for everything you guys do.
kina virus released to fk up the elections...millions dead, billion+ with heart/organ issues...evil know no bounds, 100% intentionally released...just to get one man out of office, one man they could not buy/control....Godless people will always create hell !!!
The app onto 510 E is one of my favorite processors of all time. It was a 45 white quad core laptop processor in a desktop form factor, what made it special, was that you can put it in an HTPC, NAS, or media server and it runs super quiet and super cool and has an amazing amount of power forwards TDP I was still using it up until a few months ago when the motherboard blew the USB and I haven’t found a replacement yet.
Honestly, if we are primarily comparing 7900 to 7900X, I would've liked to see 7900X PBO results in the charts as well. Because although I get the point "you can get 7900X performance by just putting 7900 on PBO", I don't think it's completely fair without including 7900X PBO as well. Not sure why that isn't included, is there something obvious I'm missing?
While this is true, scientifically speaking, GN made a video some time ago reviewing PBO on many AMD CPUs. The take-home message was: PBO almost does nothing on CPUs where thermals are not a extremely hard constraint. You may have seen the other non-X Ryzen tests, where PBO doesn't change performance (at most a few percents), basically because they were reduced from TDP 105W to 65W, the drop in thermal constraint was not even sufficient to show any difference using PBO. On the 7900 though, the performance drop from TDP 170W to 65W is big enough to notice a difference when PBO is turned on. So when the CPU is thermally unconstrained from the beginning, GN noticed that it really doesn't make any difference. That's probably why the Ryzen X-series is not tested with PBO.
Google the basic stuff ffs! 7900x is already PBO enabled. It's only constrained by temp which is how PBO OC was always done.
@@lareponse42 thanks for the info, might have missed that video!
@@alexusman list your source before you ask others to Google
@@kumarhiranya my source is brain, but it's a rare thing nowadays, I know. All X parts are only temp limited and don't reach any power constraints if there are any, that's how PBO OC is done when people mean PBO.
Great review. I really wish you would add the 13th Gen parts limited to a similar "ECO" Wattage in the BIOS so we can see how they perform in a similar setup. I know der8auer did some of this testing but would love to see it in your extensive charts
I wish the board manufacturers create a profile like that on the boards. I don't know a proper way to set those limits in the BIOS for MSI.
@@maxslayer11 its pretty straight forward for all BIOS'. I don't have an MSI so not sure how they handle it. But you can probably look it up on youtube for the Z690 boards from MSI.
Would love to see some of the benchmarks include stuff from a couple generations further back as context, I'm currently running a i9-7900x and would love to see how it compares to modern tests. Great work though keep it up!
the ryzen 9 7900 is gonna be mutch faster at everything and draw less power / get less hot
Would be really interesting to see how these non x chip perform using low profile coolers. Nice to see products that doesn't push CPU cooling to the limit in sff builds.
7600mand 7700 just fine, bigger ones could need bigger cooler if you use PBO.
I bought a Ryzen 9 7900 a couple of weeks ago as an upgrade from a Ryzen 7 5700G. My PC runs whisper quiet on using the very same CPU fan. And in a compact mATX system too.
Isn't it worth keeping the i7-13700k and i5-13600k for the 0.1% frame discussion alone? I mean in Far Cry 6. Going from 44 to 86, 0.1% is very significant. Why is frame consistency not talked about here. Isn't this important more than 190 vs 180 fps of the i9 vs i5
It would be interesting to include one 10 year old setup in the test charts as a reference, like the piledriver 8350 or sandybridge i7-3820.(PCIe2 and DDR3) For those of us that still aren't convinced hardware has advanced enough to economically justify an upgrade.
@@jbrone1241 Depends on the game of course, but the cost angle starts to become most amusing, eg. a XEON 1620 v3 (faster than the i7 3820) is only $8 but runs surprisingly well, likewise a 3930K costs diddly (18 UKP in the UK from CEX), while moving upwards even a used 6850K is still cheaper than a new i3 10100F. Alas in every csae it's finding a decent mbd which is difficult.
MyTech, there's plenty of existing data to show the differences, but for someone who supposedly cares so much about cost that they're still using decade old hw, it would make far more sense to consider used previous gen parts anyway, especially something like a used AM4 + Zen2 combo (the 2700X often sells very cheaply for some reason, at least compared to Zen3 parts). Or for core count one can go the cheap Aliexpress/XEON route, in which case Haswell XEONs become very appealing. There are so many options if price is a prime consideration.
Btw, the i7 3820 is technically SB-E, not SB, as it's for S2011 rather than S1155, though what's nice about the 3820 is it does have 40 PCIe lanes like all SB-E/IB-E CPUs, though in that regard my favourite is the unlocked 4820K (higher thermal budget per core for oc'ing, supports PCIe 3.0 native). Lastly, note that one could use an NVIDIA tool to force SB-E parts to run with PCIe 3.0, I used it with no issues. IB-E runs with 3.0 anyway, but the 4930K/4960X don't oc as easily as SB-E (there's a 5GHz Owners Club for the 3820 on OCN).
Steve when will you test idle power consumption, and gaming/ST? That's where 99% of people, especially your viewers, actually use these CPUs. Only a small fraction of people do 100% loads for extended periods.
We already do. The problem is that idle power consumption is generally so low that it amounts to noise in the measurement -- there are so many random background tasks that it's very difficult to get a true and accurate "idle" reading. Because of that, we don't present them because they are not particularly useful.
@@GamersNexus What does it mean that the idle power is noise? Zen 2/3 pulls 30 watts idling and im curious if zen 4 improves this
@@sowa705 lots of fluctuations in the power draw due to random background tasks starting, so you cant do a consistent test.
@@RexMan04 Its still good to know the approximate power draw with random things in the background.
@@sowa705 But is would be useless/misleading when comparing since it's not replicable. You would be comparing apples to oranges, that at best is just bad science and at worst it can be hugely misleading and GN would be spreading "false" information overall.
I really like your new background you've set up. Very cool :)
what about turning PBO on on the 7900X? shouldn't you get more performance if you have a good cooling solution?
That's pretty impressive. Thanks GN 👍
It's really nice to see parts that don't consume an insane amount of power
I really want to see more emphasis put into the power consumption part. I know this is a "more is better" (gamers) channel, but for the stock 7900 price you get a cooler included and for a very SFF machine its unbeatable. I really want AMD to push the power limits this way for the new generation APUs, if you get amazing 6-8 core performance/watt, they can use the rest of the available power for RDNA3 iGPU cores. Imagine a SFF machine with an APU of ~95-105W (or actually more) with Graphics performance of a 3050! (fingers crossed, assuming 40-60W GPU). That is 1080p gaming without a dedicated GPU and on a
And how much is 7900x PBO mode on then 7900 PBO mode?
Thank you Steve, for all you do. They are not selling the non X version anymore in the netherlands which sucks. perhaps i can buy it from abroad and ship it home since even that is cheaper than getting the 7900X
Excited to see if they have some mobile parts coming out with these new efficiencies.