[SPON: New Customers get $25 Off ALL CPUs: micro.center/a35965 ] [SPON: "brokensilicon“ at CDKeyOffer to get Win 11 Pro for $23: www.cdkeyoffer.com/cko/Moore11 ] [SPON: Get 10% off Tasty Vite Ramen with code BROKENSILICON: bit.ly/3wKx6v1 ] [SPON: Micro Center Charlotte is OPEN: micro.center/9aed75 ] [SPON: Early Access to Micro Center Miami: micro.center/fa62a2 ] [SPON: Shop Micro Center’s Top Deal’s: micro.center/3f31fd ] [SPON: Get 10% off Tasty Vite Ramen with code BROKENSILICON: bit.ly/3wKx6v1 ] #Windows11 #backtoschoolsale #cdkeyoffer
Honestly glad to see lunar Lake and Strix Point being fairly close, and Snapdragon in the same region. This means they have to keep prices in check, assuming each has a good amount of volume and design wins. When you have 1 company dominating, you tend to get elevated pricing
6:49 the 7800X3D is 16% faster than the 7600X in that review, but it is using older data. In the 14900K review from Techspot the 7800X3D is averaging 170 FPS to the 7600X’s 133 which is a 28% performance uplift. I would expect the average review to find Ryzen 9000 to be single digit percentages slower than the 7800X3D which does track with AMD’s statement of the new generation being slower in gaming than the current king. I think the 9600X getting as close as 5% within the 7800X3D is bit too optimistic because the 7950X is 5-10% faster than the 7600X in gaming and the 9950X isn’t the 7800X3D’s equal.
It will depend on the average, and the RAM used with the 9600X. Like I said MULTIPLE TIMES too - I do not expect it to match the 7800X3D. However it is clocked 8% faster with 16% higher IPC - that's a 25% boost in apps that benefit heavily from the IPC. It should be close in SOME averages.
The performance uplift depends on which games are tested, more of the current testing shows at roughly 20% average, so if IPC translates to gaming plus slightly higher clocks, I could see within 5% barely
@@MooresLawIsDead also Im pretty sure real world power draw of the 7800x3D is usually like 50-70 watts, either Im hallucinating or thats what I see in pretty much every benchmark since its release, cant check now but Im really quite sure its basically always below 80 watts usually below 70 watts??. So I also dont expect the 9600x to be 20-30%+ more efficient than the 7800x3D, while also being only like 5% slower. We are comparing a 6 Core non 3D-Vcache vs an 8 Core 3D-Vcache here. And as OP already noted - AMD themselves said that the non x3D Ryzen 9000 CPUs wont beat the 7800x3D. iirc they clearly said that the 7800x3D will still be the fastest gaming CPU. Daniel Owen had an article in his most recent GPU news about a rumor that AMD is considering raising the TDP of the 9700x up to 120 watts to make sure the Ryzen 9000 series isnt worse in gaming than last gen as a whole. Its an article on wccftech. Dont know if its plausible but it could make sense.
@@DELTA9XTC i think it was more of raising the tdp to beat the 7800x3d since thats what everyone is complaining about. Even though the 7800x3d does use 75-90 watts (pending on overclock/undervolt etc) it runs super hot with stacked cache. so if the 9700x comes withing 3% at the same watts it should be cooler which is a w. Similar performance to an 7800x3d, similar power draw in gaming, better use for video editing/creating etc, and lower temps, well here's to hoping they leave the tdp alone and let the 9000x3d cpu's set the new gaming benchmark
@@Ttallonn I hope for the same. Let the 9700X be 65W part or release 9700 non-X as 65W TDP. And let the 9800X3D fly. Ryzen 7700X does sell well but it's not faster in games than 5800X3D. It's roughly as fast, uses similar power but performs better overall. A lot better in some scenarios plus it has an iGPU which is awesome for old games, video playback in the background, etc. That's what sells it. I don't see how 9700X vs 7800X3D would be different
Lunar lake and beyond are exciting because intel is finally leapfrogging the industry with gate all around transistors and backside power delivery. And their gpus are also leveling up a lot. They are finally competitive and that’s great to see.
It can be a confirmation of AMD's current actual plan, but pricing can still be changed at the last minute if AMD chooses to change it. It can be changed any time after the product is released as well.
IMO, the biggest barrier to VR right now is the price. After playing Alyx, I was blown away as to what's possible. As for Apple's Vision Pro, it's not full VR and way too expensive. It's chasing a market that's not there yet. Right now, I think the biggest market for VR are simulators like DCS, MS 2020 etc, until they drop in price.
Exactly, VR on PC really only makes sense if you already have a top end PC. Otherwise you are better off buying a better CPU/GPU. I bought my Pico 4 for MSFS/Rally etc. Its an amazing experience but I already have a 7800x3d/4090 system. Still if I were on a budget id drop VR first because its only like 30% of gaming I do.
Tom I keep seeing everyone(not just you) glossing over the fact that these benchmarks for intel include the ram. So for the amd 17W vs intel 15W comparison it is really more like 19W-20W vs 15W. Which is a power difference of ~20-33%, which is significant and cannot just be brushed off like maybe you can do with 17W vs 15W. It’s not apples to apples. Just thought I would bring that up, as a lot of people keep missing it. Same even with the higher wattage… gotta add 2-3W to amd’s usage to make it apples to apples comparison… which when you are saying things like “strix looks slightly more efficient”, that 2-3 watts or ~10% power difference of the memory that isn’t included can completely change that advantage to a tie.
Ok but they also gain performance from having the RAM that close to the cores so they SHOULD be lumped together as total TDP. You would have to lower the performance if you are separating the wattage of the cores without the RAM.
Look carefully again, its AMD 15W and Intel 17W in the benchmarks. So even if you account for the RAM power consumption, its basically 17W vs 17W. Performance and numbers is one thing, you have to remember that Intel is using a much more expensive 3nm node vs AMD's 4nm. Technically AMD's design is still very efficient, maybe even superior. But the key thing is the price, AMD is able to have massive supply only only because the process node is cheaper, its also because the node is more mature and have higher yields. Availability and pricing could affect the overall value of the device, i.e AMD could have more RAM or storage etc for the same price as Intel's offering.
@@skychaos87for sure. Intel is also using n3b or whatever which was the “first run” crappy node that apple used then nobody else wanted, and it sort of got discontinued no? I can’t imagine they paid all that much for it than 4nm, because I’m not sure anyone else would even buy it. Maybe I am wrong but did anyone else use this node but apple? Seems to me tsmc/intel basically said “hey nobody wants this node, but we have it already tooled out, so we can give it to you for cheap rather than having to retool for n3p, or sell the 4nm or 3nm version that other people will actually buy”. Either that or intel are idiots and bought it before it was figured out that it was a node nobody wanted.
nobody's said this yet, but there is another AMD to Lunar Lake Comparison.. with the Lunar Lake at [30w].. is actually being compared to a [54w] AMD AI 9 HX 370, but it was by someone else who also is using the same CPU, it's just not the guy listed as the source on that one. so the comparison at higher wattages and performance may be different between the 2... i assume it favors Lunar Lake though. whoever compiled that chart in the Lunar Lake Benchmark sadly didnt post the wattage for the CPU, but im guessing it could actually be 54W just like this one, due to it being similar in total score in the same benchmark. (Timespy) If anyones interested: google: intel-core-ultra-200v-lunar-lake-reportedly-hits-4100-points-in-timespy-at-30w and its the videocardz one from about 5 days ago (from today.) and you can see the chart.
@@Azureskies01 No they shouldn’t, that is not a reason. The point is the entire system will still be more efficient because there is less power to be ran longer distances and the latency is better.
Right now is AMD's time to shine in the laptop space. This is probably one of the biggest upgrades in a long time for laptop parts because of the core improvements, graphics and NPU. Zen 6 probably won't be as big of an improvement for laptop, but because of the improvements to the MCM architecture the DESKTOP parts will get another good boost for Zen 6. So AMD has RIGHT NOW to get a good hold in the laptop space and if they fail this year to capture a decent amount then AMD will probably be stuck for a few more years with not much laptop market and not much of a chance to capture marketshare with a product that SHOULD be superior to anything else out there (forgetting Apple). I HOPE AMD gets this right and captures good marketshare because laptops need competition to bring prices down. If AMD does well then this would probably be a good gen to buy a new AMD laptop and I probably will.
The problem with laptop is AMD did not simply sell the chip to OEM. they also have to contribute the money for the laptop R&D development. AMD lose laptop market because they did not want to commit on this front during Rory Read was the CEO.
AMD couldn't capture market share when Intel was struggling due to low volume, now they gotta compete with every company under the sun to get wafers out of TSMC. Intel will be insulated from all of that if they can get 20A Arrow Lake and then next year 18A Panther Lake out into the ecosystem in large volumes. If those systems also perform well AMD is pretty much right back where they've always been as far as volume.
22:00 VRAM capacity isnt the issue for APUs on desktop, its bandwidth. My 4650G has 64GB of VRAM, but its problem is its half of the 128GB 3200MT/s ECC pool used by the whole system, on a chip that is cache starved. My 8700G has 48GB(24GB of VRAM) at 8000 in 1:1 but its still slower than an RX 6400 probably because the RX 6400 gets over 120GB/s dedicateds to it, where the 780M, despite being 4nm Zen3, has to share ~90GB/s with a cache starved CPU Strix Halo is supposedly going to help with a 256 bit memory bus rather than 128 bit, but i still would like to see APUs having the same cache as desktop moving forward. Not only does it help performance on desktop, it can also reduce RAM usage on laptop making them more efficient, yes the CPU would use more power with more cache, but the total system power is reduced. And before anyone says "But the 7800X uses way more power than the 8700G" you're comparing chiplets to monolithic. What i want you to do is compare the 4700G, and 5700G both on the same board& RAM with clock speeds locked at 4Ghz and all other specs the same (drive RAM speed/timings etc) See which one uses more power, the 4700G, partially because of the 2 CCX, but also because the system RAM uses more power, and because it needs to hit the RAM more often due to 1/2 the cache of the 5700G each bit transfered off die to RAM uses 2 orders of magnitude more power than staying within cache. Edit: For those that would like to parrot marketing excuses about not adding more cache. I'd like you to look at the RX 7600XT, and ask yourself what would cost more a 10% larger die on the 8700G for a total of 200-220mm², or the 8GB of RAM, VRM, board, licenced ports, and high end cooler included on the RX 7600XT for the same MSRP, but currently sells for less than the 8700G Oh, and did i mention the 7600XT despite being only 200mm like my proposed 32MB 8700XG, happens to come with.......32MB of cache
The problem is that cache is costly in terms of die size, since it doesn't scale well with shrinking transistor nodes. The only way forward is stacked cache, which atm is still a significant cost adder bc of the additional packaging complexity. With any luck that cost will go down over time as more advanced packaging facilities get up and running to accommodate megachips like mi300
Its the only reason I care about this apu. The 8 core 20cu version would be sick. Strix point has to many cpu cores and not enough CUs for me to care to upgrade. And I give zero fks about the "ai" shite its trying to push.
Iam waiting for the day when the intro will be starting with: And i am joined today with "super supportive sidekick...Dan" 😂 This duo is my favourite, there should be a catchy phrase
Pertaining to VR, I have a HP Reverb G2 V2 and I love it, probably would’ve gotten a PSVR2 if it was PC compatible at launch. It’s definitely a niche product. I main Driving Sim titles and I believe cost of entry is the main thing holding it back. The only VR games that feel realistic and worth it are games that you sit and play, and driving sims and flight sims are mainly it, with the cost of entry for sim hardware and a system that can run the headset at high resolution effectively being pretty high. If you want to play VR game with shooting and movement, I think the next big wave of demand will be when the “Ready Player One” hardware with haptics and the full kit is affordable including entry level gpus being able to run games with high fidelity
I played VR mini golf with my friend yesterday. We don't usually have time to see each other, so games like that are a pretty good way to socialize, better than playing battlefield and talking on discord for sure.
I bought my used 3090 more than a year and a half ago for just over $700, and I guess I could sell it for about the same amount today. That's pretty funny.
I would like to point out that a possible reason for the 2 week delay on Strix point is physical shipping constraints that couldn't be predicted when release plans were made. Specifically I'm talking about the Houthi effectively closing the Suez canal to container ships.
The metaverse will never be a reality until we figure out glasses free VR. Either a way to jack VR straight into our brains or a holodeck. We are working on both!
Regarding efficiency: The 7700X loses just 5% performance when going from 142W down to 88W PPT in Eco mode. So yeah, assuming it's still 88W PPT for 65W TDP, the 9700X will easily beat the 7700X the by around 10% due to the 16% higher IPC. That's *without* talking about any efficiency tweaks. What remains to be seen is if the change from N5 to N4 adds any efficiency gains on top of that, or there might be other currently undisclosed efficiency tweaks.
The funny part about the X Elite, when it comes to gaming, is that the M3 is not only faster for native apps but also apps running through Rosetta and GPTK!
1:16:00 Can confirm, was very surprized of the performance of the 7840HS with 4K60 4:2:2 200Mbps biggest problem is that this tiny machine (at the time) could only do 32GB of 5600MT/s and it really needed 48-64GB and should have benefitted signifiantly from 8000, as the 8700G with 48GB of 8000 is quite alot more powerful, though that may be the additional power budget
If the 3D V Cache is on both CCDs does that mean that the V Caches would be a mirror image of each other? If they aren't, would you not still have a situation where some cores may need to communicate with the V Cache on the other CCD? This would introduce latency and make it so that the 9800X3D might still be better or the same as the 9950X3D in gaming, no?
I dont care about comparing Lunar Lake to any AMD CPU nor am i worried about its performance (not everyone wants to use their laptop for gaming, and modern CPU performance is bonkers OP for the overwhelming majority of tasks). What i care about is performance per watt in comparison to the X Elite CPUs and Apple's M CPUs.
@@philipreininger2549it can't . Because lunar lake has turn off capability for cores which are unused. Amd doesn't have it. So low power and professional PC will be ruled by lunar lake only
@@AB-fc8ioI don’t believe it will have a more power efficient mode like lunar like. So if lunar lake only uses their IO die and the low-power-e-cores I think it WILL have (way) better battery life. It all depends on how strong the LP-E-Cores are, cuz if they can handle TH-cam or Word Processing they will go hard. What I do think is that in normal/slightly more intensive use cases (6-15w entire system usage) Amd will outperform Intels lunar lake. For 2 reasons, 1. if Intel uses more tiles the interconnect has some overhead like all chiplet connections do, 2. amd will have more cores with higher IPC, so power normalized the 8 or 12 cores from strix apus will run at the same clockspeed as lunar lake, use around the same power but due to them being more powerful on average the overall system will be faster, or clock lower and then be more efficient. One big plus point Intel has are fully integrated SOCs, where ram is on package meaning this saves 1W on ram and other communications which is way more costly than 0.1 to 0.4W spent on chiplet interconnect communication… so maybe Intel will be more efficient… what’s gonna be really interesting is the timing though, cuz even if Intel beats strix in a few months in efficiency, zen 6 sounds like another crazy leap next year
@@philipreininger2549 amd is also using chiplet design. Another main benefit of lunar lake is they stopped using hyper threading. It was very power Hungry. Amd will loose performance per watt game. But for hard core gamers amd will be to go chip.
0:17:55 To quote an amazon review I read months ago "Friends don't let friends buy laptops over $300." At the time I purchased an used Lenovo Thinkpad T480 for $200, It has been wonderful for carrying around to do computing and very light gaming.
It depends on what you want it for. A replacement for a desktop is a very different market segment than a portable adjunct to one. Budget DTRs - low end gaming laptops - are garbage
I bought a $200 laptop off ebay. ryzen 5700 15w part 15gb ram 256gb nvme. I will run 5 VMs and be SSHed into 3 other machines, have teams in video conference, be writing a doc and have 20 tabs in Firefox with music playing. it's fine. not a hiccup. for non 3d tasks, laptops have been great for years. I whole heartedly agree on not spending more than $300.
VR just isn't convenient enough. It's a cool gimmick... but the average consumer just doesn't want to game in VR for 6-8 hours on a Saturday. VR is multiple decades away from being something that is regularly used because the technology just isn't there. I'm sure there will be great VR experiences to be had especially for things like theme parks and laser tag type businesses. But in home use of VR will be single digit percentages of overall game time for decades to come... which does not justify the cost.
Yeah there’s not much demand for VR due to so few people using it, and since so few people use it, devs are not really incentivised to develop great VR games, which means less games coming for VR which means very few people will invest in VR due to the lack of games.
I tried it at Sandbox VR the other day. Very expensive experience, quite cool technology, game designed by a psychopath who hasn't played any proper games. The lack of killer apps & games holds back the glasses/headsets from being adopted but all the companies that make them could also afford to create the killer apps/games (or get games companies to do VR versions of existing games that make sense). For me, I want to be able to watch films in glorious 8K or something, or even really good true 4K HDR and also use it for my office PC work so I effectively have a giant monitor. It doesnt' need to be battery powered (at least, not in the headset) but it does need to last all day or have swappable battery packs. I agree we're no-where near getting the level of tech required to make visual fidelity what we want, and the headsets are very heavy currently.
Quests are cheap lmao and not that heavy. I do sim racing in vr for multiple hours a day, My track helmet puts more strain on my neck than my quest 2 does.
Virtual Desktop makes it pretty convenient on my Quest 3. 3 floating monitors to position and size is an amazing experience. VR/XR is definitely at a spot that is surprising people that get to try a good setup.
My first PC was a dell optiplex 855, my next PC was a dell XPS with an i7-880 and an HD5830. My third PC i built in 2018 has a Ryzen 2600X and an RX590. My current PC has a 3950X and a 6750XT. My next PC will be either a 9800X3D and a 8800XT or a Zen 6/RDNA 5 system! One of my buddies that works at AMD told me Medusa might have 48 CU’s, so a 20% increase plus whatever speeds and “IPC” improvements that are made to the uArch. If it’s clocked 5-8% faster, has 20% more CU’s and a +10% IPC improvement that will be a 40% improvement.
The 7800x3d' has a (relatively) higher TDP due to thermal reasons concerning the added cache. It pulls about 90 watts and basically acts as a 65W TDP model - it's virtually never hitting the PPT ceiling.
A 9950X3D seems like it could be a real monster for gaming, if it does have that cache on both the CCD's. Expensive, but really pushing gaming to the next level for CPU's. Thumbs up!
It's really unfair for FSR 3.1 reviews because the reviewers are comparing FSR 3.1 medium quality to high quality dlss because of the equal frames per second. What they should be doing is comparing Quality to Quality which you will find FSR 3.1 to be pretty much equal in high quality to dlss high quality and then compare the fps on cards of equal price to see what the upscale frames per second difference is then you'll see that they are pretty even. Only heavy ray tracing gives NVIDIA the advantage now and only a 4090 can attempt doing that with DLSS.
The info on strix halo being 3nm is cool and hopefully true. Kinda odd Tom didn't make a seporate video about it before. I kinda think strix point benifits more from a node shrink to 3nm.
If Strix Point doesn't have the AVX-512 extensions, does that mean that Strix Halo won't either? I was really looking forward to Halo to play around with the possibilities of all that bandwidth combined with serious number-crunching horsepower.
@@philipps1158 I suppose I could have been more clear. I was referring to the *improvements* on AVX-512 that the desktop Zen 5 parts are getting (as per AMD's recent presentation).
@@deansmits006bottom line is.. there are various sources and claims that suggest that zenx3d will have higher clock, architecture designed to benefit from cache even more as well as optimizations on the vcache. zen3 to zen3x3d and zen4 to zen4x3d was around +15%.. zen5x3d will be 20-25% over vanilla.. NOT simply from clock speeds.
Wonder if the Snapdragon Elit X buyers get home and install linux on it. Nice linux ARM development machine. Suggest cross compiling for android would be significantly faster than cross compiling from x64 to ARM.
This makes it difficult to choose when the best value time to buy a new laptop/handheld is...worth waiting for Zen 6? Buy Strix now in a small and light device? How about Strix Halo? Additionally is memory bandwith going to be a limitng factor for the APU's?
I'd love to go to Microcenter if there was one within a reasonable driving distance of me... there isn't, unless I'm willing to make an overnight trip about it (I'm not). Microcenter, you have my vote for an Omaha, NE location. As to CPUs, I'm very much looking forward to what the Zen 5 X3D parts bring. I'd also be open to seeing AM6 at some point in the future with a much larger socket, so that APUs with (what is now) mid-range graphics could be part of it.. add CXL for VRAM expandability so that the amount of memory on GPUs becomes a non-issue, and AMD would have a compelling advantage for their platform!
I think AMD's biggest problem is their previous gen is so frankly OP for the average person that most won't switch until their previous gen machines break, I know I can't find enough work to keep my 5700x fed and it isn't even overclocked. The software just hasn't kept up with the hardware so until I find games or programs that can make use of more than 16 threads I'll just stick with what I have.
I have the 5800x 3d and will probably continue to use it it until I need more than 64gb of ram and move back into the Intel platform. My next build will have 128gb of ram or more.
I bought my msi laptop for 870$ new on clearance at microcenter in 2022 it has an amd 8xxx cpu and rtx 3080 gpu and a 90whr battery sure bat life with the dgpu but a quick reboot gets me back to amd integrated and can easily get 12hrs
How are the prices of the 7800X3D in USA? In Europe since July 1st the 7800X3D has undergone strong price increases and is sold out at the major European distributors. Retailers increase the price of the 7800X3d several times a day.
I'm really hoping this is the start of big APUs from AMD that allow us to stop recommending GPUs like 8GB 4060s or 6500 XTs, etc. to people getting into PC gaming on strict budgets. One budget friendly chip and you're done, and see if you like PC gaming and want to go midrange or up with discrete later.
They can flood the market with new x3d chips and explode because every x3d model is proven value without the buying "fear factor" of purchasing. People that are interested won't have to "wait" for reviews to see if it's going to be a great option.
Maybe VR will take off when multimodal generative AI gets good enough to generate interesting VR environments in real-time... something as detailed as the Sora generated movie examples. Looks like AI processing capability will have to get boosted first.
Interesting to hear about the downturn in VR, I only just got into it after I heard Sony will bring compatibility to PC later this year. Hopefully Sony does not end up abandoning the VR team again.
Polaris was super popular. The same way the ATI 4000 series was super popular. It was affordable. Heres hoping they price the RX 8700/8800 XT correctly. Fuck just realised the 8800XT might take the crown as the best value 8800 (8800GT). 😅
If it’s 7900xt perf at $499 I’ll buy it day one, it’s that simple amd. I don’t want any shenanigans though or I’ll also simply stop buying amd, rdna4 will be my turning point if I go back to nvidia.
Just a short sidenote....we use lenovo t14 notebooks in business. My unit got an 8840U (Ryzen 7). The thing is going to last 5-6h productive on battery (4k display). I would argue thats pretty enough for taking notes.... ? I have basicly around 25-40 chrome tabs open...teams and outlook...vpn client and shit...i would say its not idle during my time with it. I finf this pretty impressive actually.
The classic wccftech "Reporting" regurgitating Mlid and not sourcing him. To be fair Hassan probably did that in order to not get bashed in the comments if he used you as the source. If/when it ends up being correct, he can take 'credit' for the leak and show his 'reporting' on it. If it ends up inaccurate, he will just circle back and claim it was based on Mlid, who is at fault for the erroneous pricing.
I wanted VR but two things stopped me. The price and the fact they were pumping out better units every few months, didn't want to jump in when something better was 3-4 months away.
That entry/mid level crammed into a NUC-style machine (even with a little markup for the form-factor) is something I would buy. Qualcomm I don't think has a bad product, I think they have a bad price. I would consider using ARM in said NUC-style machine if it were priced appropriately, as it's running Linux anyway.
I agree that data centres in a few years will flood the market with interesting purchasable hardware options. Service providers want you to subscribe to their software/service that uses your chips and hardware rather than finance rackspace. I feel we are seeing the start of this now.
actually lunar lake only catch up to i3 level performance but it cost very high for laptop manufacturer, due to memory on package, TSMC N3E tech, difference package.
27:09 most people I know who want to go this route, with a " starter APU" aren't truly trying to save money. What they're trying to do is extend the purchasing timeline while still having a perfectly workable system that at least let's them enjoy it. So for example they really only have a $1,000 to spend today, but will have another 1,000 in a month or two. Or if they are trying to save money, it's because they're trying to buy piecemeal on sale certain items. Like the case, power supply and motherboard/memory at one point in time, knowing there will be probably sales or a new version or better version of the CPU later, and the same with the GPU. Because it's only a few months before the new release comes out but they would like to have something for the summer... Envisioning their actual end state being in about 6 months
AMD has had deals with ASUS since the white box Athlon 500\550 bundle... They had the first AMD Advantage laptops (and mostly only)... That's why Strix Halo is a necessary product... 4070 perf in a single chip solution...? They can cut Nvidia out of more laptops since why put a 4060\5060 in when you can save on cooling and complexity... I can only assume that the usual suspects will make killer MiniPCs with it next year...
Go buy the house burning i9-14900K, and then don't cry to me when you need to RMA it. I'm not gonna make bad recommendations just to be "fair". And clearly, you have no memory of this channel...which is basically the only one that leaked dozens of Intel products and was hyped for ARC until it was clear it would fail (and it did btw).
"Secret covid spreader" Now that's funny xD I just went out like nothing had changed. Got it 4 times, emergency room once and laid out for two weeks doing nothing the second. You saw me out in public and you knew I had it LOL
@@rainbowbunchie8237 I am in a different customers home every day so I am subjected to all manner of germs daily. Been doing it for 20 years. I'm sure she is happy about it.
I could def see Zen 5's being almost half the power at same performance, Zen 4 is using N5, Zen 5 is using N4X, which is the 5th iteration of TSMC's 5nm process, it should be much better
@@benpedi6645 did u read my comment properly, i'm saying that Zen 5 should consume almost half the power for same performance (Edit: in tom's scenario, altou 7800x3d almost never runs at it's 120w tdp)
Any information on how the 5800xt and 5900xt might perform? I'm especially wondering if the 5800xt at $250 would perform significantly better than a 5700x3d which retails for $200 new right now since it boosts higher but doesn't have 3d v cache. Also any idea if there will be a 5800xt 3d?
I'm still surprised that AMD hasn't used their increased profits and engineers to help UE5 add FSR to the engine so that it's available as an option to developers. Even more surprised that Nividia did do that a long time ago. Maybe now that AMD is making it be in separate coding libraries, they will work with UE. And I'm talking about a built-in option not a requirement, still leave it up to the actual game developers.
Should i wait for Ryzen AI HX 370 based G14 or G16 or go ahead and get the 8945HS Oled G14, is this AI NPU power helpful in the long run? It already gas a 4060 or 4070, so doesnt matter if its 780m or 890m
AMD's choice is to use their TSM advanced packaging allocation for $400 consumer chips vs $20,000 AI chips. Let me guess how they will use that allocation ...
The Problem with the ZEN5 Efficiency thing is that a 7700X loses 2%, yes two percent performance going from 105W TDP (142W PPT) to 65W TDP (88W PPT) so when somebody from AMD says that ZEN5 8 Core doesn't need 105W TDP, that says nothing about it's Efficiency. ZEN4 X SKUs were just pushed without reason.
@@Games_and_Tech Nah, for 2 % it's just silly, also the direct Competitor (13600K) was/is still faster in MT. For 7950X it kinda made sense to use 230W instead of 142W, because it gave 5% performance and that made it slightly beat 13900K in MT. Still silly, but not absolutely pointless like 7700X.
First thing i do when i get a new CPU is check the headroom for fun and find that there is not much as there at there limit, then i pull back the voltage to find the sweet spot for perf and thermals, and im good to go, 2% ipc may not even be a fram or two.
i mean i understand why they didnt disclose efficiency on the strix halo.... theres nothing to compare it to. the 8000G apus are gonna be blown away in the graphics department by the strix. i would call it efficient in any context because you dont need a gpu for low end gaming
The killer feature of these APUs, compared to the same expenditure in CPU + cheap old GPU, is power efficiency. At European electricity prices, an extra 100W in 8 hour/day use can burn $50-$100 in a year. Add that up over a few years, and the APU makes a LOT more sense than the CPU + old GPU option. Also, cost aside, burning electricity 'round here is still contributing to Putin's war chest, and of course it isn't much help to the planet either. I'm rocking a $200 8600g for now and it's just fine.
At 1hr into the video, during the X3d discussion it's mentioned that X3d is launching at the same time as non X3d.. Wait, what?!?!? I'm pretty sure that this is not the case.
It is four weeks (previously two) to Strix Point launching. Asus has three models announced (S 16, ProArt 16, G16). There's some Geekbench results for an unnamed Lenovo model. Acer has a Swift 14 that also just showed up in tests. That's not a lot of models, even compared to X Elite. Moreover there are just two APUs to choose from, again even less than X Elite. This doesn't really feel like a launch with massive volume.
True. However, it's important to point out several skus is good for OEM and manufacturers, but not in anyway desirable for consumers. Two is plenty. Six differents X chip is a drag.
The fact that so many people are foretelling of doom because AMD looks to only be releasing a 70 class card this gen really shows how powerful the halo effect is in the GPU market. It's like thinking every car manufacturer is giving up and going home because they aren't competing with Bugatti's latest hypercar. Just because something doesn't compete at the absolute top of the market doesn't mean it isn't a fantastic product in its own right.
@@djchristian82 not really. The architecture work for rdna4 was already done. The real reason it was cancelled is that consumer graphics has the lowest margins in entire amd and the very complex multi chip module high end rdna 4 gpu woud have competed with limited packaging capacity with high margin data center products.
I'm starting to notice prices dip a bit. I saw an INTEL 13700 with NVDIA 3060 GPU (the ones with 12 gigs VRAM) for under $900. I was tempted to buy it, but I stood my ground and will wait a few more months. I still think prices are going to come down a bit more 🙂 I don't need a top of the line computer. I just want one that will be good for 3 or 4 years before I buy another. Usually you can do that with an $800 to $1000 computer when they are on sale.
I can corroborate the claim of late supply for X elite models - down here in Singapore we only received snapdragon laptops *this week* for retail shelves. Just Asus S15 models
I'm not a gamer but I watch this 'cause I love tech, and, I'm watching this in VR lol. I work at home in a cramped little room, and I much prefer working in a Virtual Desktop environment that supports dual monitors now and where I can switch seats and watch TH-cam on the big screen in my fancy VR apartment where I can look out the big windows and see the cars and people moving about in the streets below. I also like Immersed and I'm interested in their Visor, but Virtual Desktop is less friction for when I want to put on my Quest Pro and use my treadmill as a standing desk or for slow walking when working and fast walking when watching stuff on the big screen. The Quest 3 can support 3 monitors in Virtual Desktop with much better resolution. It helps me because I can get exercise while working, don't need more monitors for that, and can feel less cramped and can focus better on coding. If someone calls me on Teams I can just put down the headset (when using Virtual Desktop) and sit down at my monitors. If all my team were to work in Immersed that would be viable too. I say I'm not a gamer, but I do like Microsoft Flight Simulator in VR which is meant to be better supported in the new release later this year, and I do like particular applications like the kayaking simulator thing. I tend to play those using Steam VR with Virtual Desktop. I use my desktop's WiFi card as a hotspot. If I had space I would like things like beat-sabre and virtual table tennis etc. but I don't have the space. Anyway. VR is important for me. Very important! It's what loosens my grip on money lol. Though I do want the RTX 5090 for AI lol. I guess I'm very much a minority. I must be an "alien" like Mark Z. :)
nb: I went from moving from desktop's to mini-pcs / NUCs etc. which were great for Proxmox and all the IT and Dev requirements I had, to a desktop again ... I7 Raptor Lake and RTX 4070, just because of AI as GPT 4 came out last year, and with VR being a plus. I bought into it then. And my enthusiasm has only increased as VR and AI do deliver for me. I realise AI, as per my understanding and experience with it so far, will likely completely replace Devs eventually with a number of job-transitions leading up to that in the shorter to mid-term, but I'm still enthusiastic for that future even with the surety it will delete my job and my main vocational skillsets. I'm passionate about tech - actually working on projects ... hardware and software and IT etc. These innovations provide solutions to the projects I'm working on - enabling and facilitating. I'll just need a new source of income if Universal Income doesn't become a thing. "AI" models and architectures improve all the time ... already I can do more "AI" stuff on my 4070 now than I could a year ago ... and I hope the same trend will continue with the 5090. It will do a lot for me in the short-term, and a lot more for me in the long term.
@@Rojfos that sounds rather like apple does not care. It's not EU which prevented them, it's just apple which was neglectant to produce enough to supply the other market which is as big as the US. Or EU was Plan B incase they can't sell enough in the US
[SPON: New Customers get $25 Off ALL CPUs: micro.center/a35965 ]
[SPON: "brokensilicon“ at CDKeyOffer to get Win 11 Pro for $23: www.cdkeyoffer.com/cko/Moore11 ]
[SPON: Get 10% off Tasty Vite Ramen with code BROKENSILICON: bit.ly/3wKx6v1 ]
[SPON: Micro Center Charlotte is OPEN: micro.center/9aed75 ]
[SPON: Early Access to Micro Center Miami: micro.center/fa62a2 ]
[SPON: Shop Micro Center’s Top Deal’s: micro.center/3f31fd ]
[SPON: Get 10% off Tasty Vite Ramen with code BROKENSILICON: bit.ly/3wKx6v1 ]
#Windows11 #backtoschoolsale #cdkeyoffer
Fair comparison: Radeon 890M at 54W(you hide it) vs LNL Arc graphics at 30W!
Good Job AMD fan! 👏 😂
Honestly glad to see lunar Lake and Strix Point being fairly close, and Snapdragon in the same region. This means they have to keep prices in check, assuming each has a good amount of volume and design wins. When you have 1 company dominating, you tend to get elevated pricing
6:49 the 7800X3D is 16% faster than the 7600X in that review, but it is using older data. In the 14900K review from Techspot the 7800X3D is averaging 170 FPS to the 7600X’s 133 which is a 28% performance uplift. I would expect the average review to find Ryzen 9000 to be single digit percentages slower than the 7800X3D which does track with AMD’s statement of the new generation being slower in gaming than the current king. I think the 9600X getting as close as 5% within the 7800X3D is bit too optimistic because the 7950X is 5-10% faster than the 7600X in gaming and the 9950X isn’t the 7800X3D’s equal.
It will depend on the average, and the RAM used with the 9600X. Like I said MULTIPLE TIMES too - I do not expect it to match the 7800X3D. However it is clocked 8% faster with 16% higher IPC - that's a 25% boost in apps that benefit heavily from the IPC. It should be close in SOME averages.
The performance uplift depends on which games are tested, more of the current testing shows at roughly 20% average, so if IPC translates to gaming plus slightly higher clocks, I could see within 5% barely
@@MooresLawIsDead also Im pretty sure real world power draw of the 7800x3D is usually like 50-70 watts, either Im hallucinating or thats what I see in pretty much every benchmark since its release, cant check now but Im really quite sure its basically always below 80 watts usually below 70 watts??. So I also dont expect the 9600x to be 20-30%+ more efficient than the 7800x3D, while also being only like 5% slower. We are comparing a 6 Core non 3D-Vcache vs an 8 Core 3D-Vcache here. And as OP already noted - AMD themselves said that the non x3D Ryzen 9000 CPUs wont beat the 7800x3D. iirc they clearly said that the 7800x3D will still be the fastest gaming CPU.
Daniel Owen had an article in his most recent GPU news about a rumor that AMD is considering raising the TDP of the 9700x up to 120 watts to make sure the Ryzen 9000 series isnt worse in gaming than last gen as a whole. Its an article on wccftech. Dont know if its plausible but it could make sense.
@@DELTA9XTC i think it was more of raising the tdp to beat the 7800x3d since thats what everyone is complaining about. Even though the 7800x3d does use 75-90 watts (pending on overclock/undervolt etc) it runs super hot with stacked cache. so if the 9700x comes withing 3% at the same watts it should be cooler which is a w. Similar performance to an 7800x3d, similar power draw in gaming, better use for video editing/creating etc, and lower temps, well here's to hoping they leave the tdp alone and let the 9000x3d cpu's set the new gaming benchmark
@@Ttallonn I hope for the same. Let the 9700X be 65W part or release 9700 non-X as 65W TDP. And let the 9800X3D fly.
Ryzen 7700X does sell well but it's not faster in games than 5800X3D. It's roughly as fast, uses similar power but performs better overall. A lot better in some scenarios plus it has an iGPU which is awesome for old games, video playback in the background, etc. That's what sells it. I don't see how 9700X vs 7800X3D would be different
Lunar lake and beyond are exciting because intel is finally leapfrogging the industry with gate all around transistors and backside power delivery. And their gpus are also leveling up a lot. They are finally competitive and that’s great to see.
People always say, Never put “Confirmed” in the title
May be Tom wants to take a chance
Its called clickbait
Can you confirm which people?
It can be a confirmation of AMD's current actual plan, but pricing can still be changed at the last minute if AMD chooses to change it. It can be changed any time after the product is released as well.
Does anyone have a timestamp?
IMO, the biggest barrier to VR right now is the price. After playing Alyx, I was blown away as to what's possible. As for Apple's Vision Pro, it's not full VR and way too expensive. It's chasing a market that's not there yet. Right now, I think the biggest market for VR are simulators like DCS, MS 2020 etc, until they drop in price.
Exactly, VR on PC really only makes sense if you already have a top end PC. Otherwise you are better off buying a better CPU/GPU. I bought my Pico 4 for MSFS/Rally etc. Its an amazing experience but I already have a 7800x3d/4090 system. Still if I were on a budget id drop VR first because its only like 30% of gaming I do.
The biggest barrier to vr isn't price. It's play space.
Tom I keep seeing everyone(not just you) glossing over the fact that these benchmarks for intel include the ram. So for the amd 17W vs intel 15W comparison it is really more like 19W-20W vs 15W. Which is a power difference of ~20-33%, which is significant and cannot just be brushed off like maybe you can do with 17W vs 15W. It’s not apples to apples. Just thought I would bring that up, as a lot of people keep missing it. Same even with the higher wattage… gotta add 2-3W to amd’s usage to make it apples to apples comparison… which when you are saying things like “strix looks slightly more efficient”, that 2-3 watts or ~10% power difference of the memory that isn’t included can completely change that advantage to a tie.
Ok but they also gain performance from having the RAM that close to the cores so they SHOULD be lumped together as total TDP. You would have to lower the performance if you are separating the wattage of the cores without the RAM.
Look carefully again, its AMD 15W and Intel 17W in the benchmarks. So even if you account for the RAM power consumption, its basically 17W vs 17W. Performance and numbers is one thing, you have to remember that Intel is using a much more expensive 3nm node vs AMD's 4nm. Technically AMD's design is still very efficient, maybe even superior. But the key thing is the price, AMD is able to have massive supply only only because the process node is cheaper, its also because the node is more mature and have higher yields. Availability and pricing could affect the overall value of the device, i.e AMD could have more RAM or storage etc for the same price as Intel's offering.
@@skychaos87for sure. Intel is also using n3b or whatever which was the “first run” crappy node that apple used then nobody else wanted, and it sort of got discontinued no? I can’t imagine they paid all that much for it than 4nm, because I’m not sure anyone else would even buy it. Maybe I am wrong but did anyone else use this node but apple? Seems to me tsmc/intel basically said “hey nobody wants this node, but we have it already tooled out, so we can give it to you for cheap rather than having to retool for n3p, or sell the 4nm or 3nm version that other people will actually buy”. Either that or intel are idiots and bought it before it was figured out that it was a node nobody wanted.
nobody's said this yet, but there is another AMD to Lunar Lake Comparison.. with the Lunar Lake at [30w].. is actually being compared to a [54w] AMD AI 9 HX 370, but it was by someone else who also is using the same CPU, it's just not the guy listed as the source on that one.
so the comparison at higher wattages and performance may be different between the 2... i assume it favors Lunar Lake though. whoever compiled that chart in the Lunar Lake Benchmark sadly didnt post the wattage for the CPU, but im guessing it could actually be 54W just like this one, due to it being similar in total score in the same benchmark. (Timespy)
If anyones interested: google: intel-core-ultra-200v-lunar-lake-reportedly-hits-4100-points-in-timespy-at-30w and its the videocardz one from about 5 days ago (from today.)
and you can see the chart.
@@Azureskies01 No they shouldn’t, that is not a reason. The point is the entire system will still be more efficient because there is less power to be ran longer distances and the latency is better.
Right now is AMD's time to shine in the laptop space. This is probably one of the biggest upgrades in a long time for laptop parts because of the core improvements, graphics and NPU.
Zen 6 probably won't be as big of an improvement for laptop, but because of the improvements to the MCM architecture the DESKTOP parts will get another good boost for Zen 6.
So AMD has RIGHT NOW to get a good hold in the laptop space and if they fail this year to capture a decent amount then AMD will probably be stuck for a few more years with not much laptop market and not much of a chance to capture marketshare with a product that SHOULD be superior to anything else out there (forgetting Apple). I HOPE AMD gets this right and captures good marketshare because laptops need competition to bring prices down. If AMD does well then this would probably be a good gen to buy a new AMD laptop and I probably will.
The problem with laptop is AMD did not simply sell the chip to OEM. they also have to contribute the money for the laptop R&D development. AMD lose laptop market because they did not want to commit on this front during Rory Read was the CEO.
AMD couldn't capture market share when Intel was struggling due to low volume, now they gotta compete with every company under the sun to get wafers out of TSMC. Intel will be insulated from all of that if they can get 20A Arrow Lake and then next year 18A Panther Lake out into the ecosystem in large volumes. If those systems also perform well AMD is pretty much right back where they've always been as far as volume.
22:00 VRAM capacity isnt the issue for APUs on desktop, its bandwidth.
My 4650G has 64GB of VRAM, but its problem is its half of the 128GB 3200MT/s ECC pool used by the whole system, on a chip that is cache starved.
My 8700G has 48GB(24GB of VRAM) at 8000 in 1:1 but its still slower than an RX 6400 probably because the RX 6400 gets over 120GB/s dedicateds to it, where the 780M, despite being 4nm Zen3, has to share ~90GB/s with a cache starved CPU
Strix Halo is supposedly going to help with a 256 bit memory bus rather than 128 bit, but i still would like to see APUs having the same cache as desktop moving forward.
Not only does it help performance on desktop, it can also reduce RAM usage on laptop making them more efficient, yes the CPU would use more power with more cache, but the total system power is reduced.
And before anyone says "But the 7800X uses way more power than the 8700G" you're comparing chiplets to monolithic. What i want you to do is compare the 4700G, and 5700G both on the same board& RAM with clock speeds locked at 4Ghz and all other specs the same (drive RAM speed/timings etc)
See which one uses more power, the 4700G, partially because of the 2 CCX, but also because the system RAM uses more power, and because it needs to hit the RAM more often due to 1/2 the cache of the 5700G each bit transfered off die to RAM uses 2 orders of magnitude more power than staying within cache.
Edit: For those that would like to parrot marketing excuses about not adding more cache.
I'd like you to look at the RX 7600XT, and ask yourself what would cost more
a 10% larger die on the 8700G for a total of 200-220mm², or the 8GB of RAM, VRM, board, licenced ports, and high end cooler included on the RX 7600XT for the same MSRP, but currently sells for less than the 8700G
Oh, and did i mention the 7600XT despite being only 200mm like my proposed 32MB 8700XG, happens to come with.......32MB of cache
APUs are least discussed......it sucks.
The problem is that cache is costly in terms of die size, since it doesn't scale well with shrinking transistor nodes. The only way forward is stacked cache, which atm is still a significant cost adder bc of the additional packaging complexity. With any luck that cost will go down over time as more advanced packaging facilities get up and running to accommodate megachips like mi300
Strix halo x3d since it is using chip let's with 32mb mall cache. Hype train
@@bobo-cc1xw May from Zen 6.......if AMD is generous enough 😅
@@APU-iGPU Lisa su be praised
*Strix Halo Handheld would be insane*
Its the only reason I care about this apu. The 8 core 20cu version would be sick. Strix point has to many cpu cores and not enough CUs for me to care to upgrade. And I give zero fks about the "ai" shite its trying to push.
Not sure if anyone else noticed but Dan looks way more crisp today😂
New camera
Speaking of people who are about to get laid...
He is opening a OnlyFans soon
@@pedro.alcatra 🤣🤣🤣
Iam waiting for the day when the intro will be starting with:
And i am joined today with "super supportive sidekick...Dan" 😂
This duo is my favourite, there should be a catchy phrase
Pertaining to VR, I have a HP Reverb G2 V2 and I love it, probably would’ve gotten a PSVR2 if it was PC compatible at launch. It’s definitely a niche product. I main Driving Sim titles and I believe cost of entry is the main thing holding it back. The only VR games that feel realistic and worth it are games that you sit and play, and driving sims and flight sims are mainly it, with the cost of entry for sim hardware and a system that can run the headset at high resolution effectively being pretty high. If you want to play VR game with shooting and movement, I think the next big wave of demand will be when the “Ready Player One” hardware with haptics and the full kit is affordable including entry level gpus being able to run games with high fidelity
I played VR mini golf with my friend yesterday. We don't usually have time to see each other, so games like that are a pretty good way to socialize, better than playing battlefield and talking on discord for sure.
I hope RDNA 4 will have vastly superior ray tracing and RDNA 5 will being both horsepower and ray tracing.
In that case, you mean more Gigahertz.
I bought my used 3090 more than a year and a half ago for just over $700, and I guess I could sell it for about the same amount today. That's pretty funny.
And get a 5090 after
@@thereddog223 I don't really need a 5090. I rarely even make full use of my 3090 honestly
@@CutoutClips Could do with less power consumption/heat/noise myself
@@CutoutClipsget a used 4090 when 50 series releases :)
I would like to point out that a possible reason for the 2 week delay on Strix point is physical shipping constraints that couldn't be predicted when release plans were made.
Specifically I'm talking about the Houthi effectively closing the Suez canal to container ships.
The metaverse will never be a reality until we figure out glasses free VR. Either a way to jack VR straight into our brains or a holodeck. We are working on both!
Regarding efficiency: The 7700X loses just 5% performance when going from 142W down to 88W PPT in Eco mode. So yeah, assuming it's still 88W PPT for 65W TDP, the 9700X will easily beat the 7700X the by around 10% due to the 16% higher IPC. That's *without* talking about any efficiency tweaks. What remains to be seen is if the change from N5 to N4 adds any efficiency gains on top of that, or there might be other currently undisclosed efficiency tweaks.
The funny part about the X Elite, when it comes to gaming, is that the M3 is not only faster for native apps but also apps running through Rosetta and GPTK!
That thing needs like 2 translation layers to compete. Overly complicated apple bullshit. At least prism is out of the box
Now qualcom needs to get in on the GPU game
1:16:00 Can confirm, was very surprized of the performance of the 7840HS with 4K60 4:2:2 200Mbps biggest problem is that this tiny machine (at the time) could only do 32GB of 5600MT/s and it really needed 48-64GB and should have benefitted signifiantly from 8000, as the 8700G with 48GB of 8000 is quite alot more powerful, though that may be the additional power budget
Man I just hope AMD doesn't repeat the mistake of announcing RDNA4 together with FSR 4 only to not launch the feature for 6+ months.
Tuesdays are my favorite! Thanks to MLID :)
Mehner Lake is the next rumored Intel release and they are finally giving realistic names to these designs ... Then to be followed by Midner Lake ...
dying
Zuckerberg read Snow Crash and realized the guys in charge of the Metaverse were multi-billionaires.
If the 3D V Cache is on both CCDs does that mean that the V Caches would be a mirror image of each other? If they aren't, would you not still have a situation where some cores may need to communicate with the V Cache on the other CCD? This would introduce latency and make it so that the 9800X3D might still be better or the same as the 9950X3D in gaming, no?
I dont care about comparing Lunar Lake to any AMD CPU nor am i worried about its performance (not everyone wants to use their laptop for gaming, and modern CPU performance is bonkers OP for the overwhelming majority of tasks). What i care about is performance per watt in comparison to the X Elite CPUs and Apple's M CPUs.
They will all be pretty comparable.
My worry is amd will be more efficient with more cores and gpu performance still
@@philipreininger2549it can't . Because lunar lake has turn off capability for cores which are unused. Amd doesn't have it. So low power and professional PC will be ruled by lunar lake only
@@AB-fc8ioI don’t believe it will have a more power efficient mode like lunar like. So if lunar lake only uses their IO die and the low-power-e-cores I think it WILL have (way) better battery life. It all depends on how strong the LP-E-Cores are, cuz if they can handle TH-cam or Word Processing they will go hard.
What I do think is that in normal/slightly more intensive use cases (6-15w entire system usage) Amd will outperform Intels lunar lake. For 2 reasons, 1. if Intel uses more tiles the interconnect has some overhead like all chiplet connections do, 2. amd will have more cores with higher IPC, so power normalized the 8 or 12 cores from strix apus will run at the same clockspeed as lunar lake, use around the same power but due to them being more powerful on average the overall system will be faster, or clock lower and then be more efficient.
One big plus point Intel has are fully integrated SOCs, where ram is on package meaning this saves 1W on ram and other communications which is way more costly than 0.1 to 0.4W spent on chiplet interconnect communication… so maybe Intel will be more efficient… what’s gonna be really interesting is the timing though, cuz even if Intel beats strix in a few months in efficiency, zen 6 sounds like another crazy leap next year
@@philipreininger2549 amd is also using chiplet design. Another main benefit of lunar lake is they stopped using hyper threading. It was very power Hungry. Amd will loose performance per watt game. But for hard core gamers amd will be to go chip.
Why do you guys look like brothers all of a sudden? When the podcast first started i had trouble figuring out who was who.
Pretty sure they are brothers.
Medusa Halo sounds like the perfect apu for tthe future steam machine. Would fit the 2027 release.
0:17:55 To quote an amazon review I read months ago "Friends don't let friends buy laptops over $300."
At the time I purchased an used Lenovo Thinkpad T480 for $200, It has been wonderful for carrying around to do computing and very light gaming.
It depends on what you want it for. A replacement for a desktop is a very different market segment than a portable adjunct to one. Budget DTRs - low end gaming laptops - are garbage
I bought a $200 laptop off ebay. ryzen 5700 15w part 15gb ram 256gb nvme. I will run 5 VMs and be SSHed into 3 other machines, have teams in video conference, be writing a doc and have 20 tabs in Firefox with music playing. it's fine. not a hiccup. for non 3d tasks, laptops have been great for years. I whole heartedly agree on not spending more than $300.
@@lilnapkin462 that is a fantastic deal, congratulations on finding that!
VR just isn't convenient enough. It's a cool gimmick... but the average consumer just doesn't want to game in VR for 6-8 hours on a Saturday. VR is multiple decades away from being something that is regularly used because the technology just isn't there. I'm sure there will be great VR experiences to be had especially for things like theme parks and laser tag type businesses. But in home use of VR will be single digit percentages of overall game time for decades to come... which does not justify the cost.
Yeah there’s not much demand for VR due to so few people using it, and since so few people use it, devs are not really incentivised to develop great VR games, which means less games coming for VR which means very few people will invest in VR due to the lack of games.
I tried it at Sandbox VR the other day. Very expensive experience, quite cool technology, game designed by a psychopath who hasn't played any proper games. The lack of killer apps & games holds back the glasses/headsets from being adopted but all the companies that make them could also afford to create the killer apps/games (or get games companies to do VR versions of existing games that make sense). For me, I want to be able to watch films in glorious 8K or something, or even really good true 4K HDR and also use it for my office PC work so I effectively have a giant monitor. It doesnt' need to be battery powered (at least, not in the headset) but it does need to last all day or have swappable battery packs. I agree we're no-where near getting the level of tech required to make visual fidelity what we want, and the headsets are very heavy currently.
Quests are cheap lmao and not that heavy. I do sim racing in vr for multiple hours a day, My track helmet puts more strain on my neck than my quest 2 does.
Virtual Desktop makes it pretty convenient on my Quest 3.
3 floating monitors to position and size is an amazing experience.
VR/XR is definitely at a spot that is surprising people that get to try a good setup.
Guessing 2030s
My first PC was a dell optiplex 855, my next PC was a dell XPS with an i7-880 and an HD5830. My third PC i built in 2018 has a Ryzen 2600X and an RX590. My current PC has a 3950X and a 6750XT. My next PC will be either a 9800X3D and a 8800XT or a Zen 6/RDNA 5 system!
One of my buddies that works at AMD told me Medusa might have 48 CU’s, so a 20% increase plus whatever speeds and “IPC” improvements that are made to the uArch. If it’s clocked 5-8% faster, has 20% more CU’s and a +10% IPC improvement that will be a 40% improvement.
The 7800x3d' has a (relatively) higher TDP due to thermal reasons concerning the added cache. It pulls about 90 watts and basically acts as a 65W TDP model - it's virtually never hitting the PPT ceiling.
A 9950X3D seems like it could be a real monster for gaming, if it does have that cache on both the CCD's. Expensive, but really pushing gaming to the next level for CPU's. Thumbs up!
Not really when 4090 not bottlenecked with 7800x3d
It's really unfair for FSR 3.1 reviews because the reviewers are comparing FSR 3.1 medium quality to high quality dlss because of the equal frames per second. What they should be doing is comparing Quality to Quality which you will find FSR 3.1 to be pretty much equal in high quality to dlss high quality and then compare the fps on cards of equal price to see what the upscale frames per second difference is then you'll see that they are pretty even. Only heavy ray tracing gives NVIDIA the advantage now and only a 4090 can attempt doing that with DLSS.
The info on strix halo being 3nm is cool and hopefully true. Kinda odd Tom didn't make a seporate video about it before. I kinda think strix point benifits more from a node shrink to 3nm.
Wasnt it mentioned in last leak video definitely something with strix halo being released later but on 3nm
41:38 Whenever Tom throws to Dan it's:
"I mean... (grimace | scratch head) ... you know ... "
Maybe if Tom lets him finish his ideas...
Stay mad noob
if 9950x3d 2 vcache ccds im sold
If Strix Point doesn't have the AVX-512 extensions, does that mean that Strix Halo won't either? I was really looking forward to Halo to play around with the possibilities of all that bandwidth combined with serious number-crunching horsepower.
They all have avx-512. Just maybe not Double it like the newest gen would
@@philipps1158 I suppose I could have been more clear. I was referring to the *improvements* on AVX-512 that the desktop Zen 5 parts are getting (as per AMD's recent presentation).
Zen5 Hitting 5.6GHz with higher IPC is still going to be pretty quick compared to 7800X3D at 5GHz.
Yes. But, how do we know Zen 5 x3d will boost that high?
@@deansmits006bottom line is.. there are various sources and claims that suggest that zenx3d will have higher clock, architecture designed to benefit from cache even more as well as optimizations on the vcache. zen3 to zen3x3d and zen4 to zen4x3d was around +15%.. zen5x3d will be 20-25% over vanilla.. NOT simply from clock speeds.
@@deansmits006**in games**
Wonder if the Snapdragon Elit X buyers get home and install linux on it. Nice linux ARM development machine. Suggest cross compiling for android would be significantly faster than cross compiling from x64 to ARM.
Give us more on zen 5 x3d
This makes it difficult to choose when the best value time to buy a new laptop/handheld is...worth waiting for Zen 6? Buy Strix now in a small and light device? How about Strix Halo? Additionally is memory bandwith going to be a limitng factor for the APU's?
Zuck is basically the guy from the "Don't Create the Torment Nexus" tweet
Tom said he ISN’T dying…. Does that mean he secretly is?
Probably means I'm a secret Nvidia Fanboy
@@MooresLawIsDeador that lol
I'd love to go to Microcenter if there was one within a reasonable driving distance of me... there isn't, unless I'm willing to make an overnight trip about it (I'm not). Microcenter, you have my vote for an Omaha, NE location. As to CPUs, I'm very much looking forward to what the Zen 5 X3D parts bring. I'd also be open to seeing AM6 at some point in the future with a much larger socket, so that APUs with (what is now) mid-range graphics could be part of it.. add CXL for VRAM expandability so that the amount of memory on GPUs becomes a non-issue, and AMD would have a compelling advantage for their platform!
I think AMD's biggest problem is their previous gen is so frankly OP for the average person that most won't switch until their previous gen machines break, I know I can't find enough work to keep my 5700x fed and it isn't even overclocked. The software just hasn't kept up with the hardware so until I find games or programs that can make use of more than 16 threads I'll just stick with what I have.
I have a 5700x too.
I have the 5800x 3d and will probably continue to use it it until I need more than 64gb of ram and move back into the Intel platform. My next build will have 128gb of ram or more.
I bought my msi laptop for 870$ new on clearance at microcenter in 2022 it has an amd 8xxx cpu and rtx 3080 gpu and a 90whr battery sure bat life with the dgpu but a quick reboot gets me back to amd integrated and can easily get 12hrs
Just copped a 4080 super for $699, did i get a steal or am i gimping myself on not just waiting for the 5000 series? (currently on 3080)
How are the prices of the 7800X3D in USA?
In Europe since July 1st the 7800X3D has undergone strong price increases and is sold out at the major European distributors. Retailers increase the price of the 7800X3d several times a day.
I'm really hoping this is the start of big APUs from AMD that allow us to stop recommending GPUs like 8GB 4060s or 6500 XTs, etc. to people getting into PC gaming on strict budgets. One budget friendly chip and you're done, and see if you like PC gaming and want to go midrange or up with discrete later.
They can flood the market with new x3d chips and explode because every x3d model is proven value without the buying "fear factor" of purchasing. People that are interested won't have to "wait" for reviews to see if it's going to be a great option.
Maybe VR will take off when multimodal generative AI gets good enough to generate interesting VR environments in real-time... something as detailed as the Sora generated movie examples. Looks like AI processing capability will have to get boosted first.
Interesting to hear about the downturn in VR, I only just got into it after I heard Sony will bring compatibility to PC later this year. Hopefully Sony does not end up abandoning the VR team again.
Polaris was super popular. The same way the ATI 4000 series was super popular. It was affordable. Heres hoping they price the RX 8700/8800 XT correctly. Fuck just realised the 8800XT might take the crown as the best value 8800 (8800GT). 😅
If it’s 7900xt perf at $499 I’ll buy it day one, it’s that simple amd. I don’t want any shenanigans though or I’ll also simply stop buying amd, rdna4 will be my turning point if I go back to nvidia.
Just a short sidenote....we use lenovo t14 notebooks in business. My unit got an 8840U (Ryzen 7). The thing is going to last 5-6h productive on battery (4k display). I would argue thats pretty enough for taking notes.... ?
I have basicly around 25-40 chrome tabs open...teams and outlook...vpn client and shit...i would say its not idle during my time with it.
I finf this pretty impressive actually.
The classic wccftech "Reporting" regurgitating Mlid and not sourcing him. To be fair Hassan probably did that in order to not get bashed in the comments if he used you as the source. If/when it ends up being correct, he can take 'credit' for the leak and show his 'reporting' on it. If it ends up inaccurate, he will just circle back and claim it was based on Mlid, who is at fault for the erroneous pricing.
I wanted VR but two things stopped me. The price and the fact they were pumping out better units every few months, didn't want to jump in when something better was 3-4 months away.
That entry/mid level crammed into a NUC-style machine (even with a little markup for the form-factor) is something I would buy.
Qualcomm I don't think has a bad product, I think they have a bad price. I would consider using ARM in said NUC-style machine if it were priced appropriately, as it's running Linux anyway.
id actually love a deal where u get an amd gpu+cpu as long as theres an actual saving
I agree that data centres in a few years will flood the market with interesting purchasable hardware options. Service providers want you to subscribe to their software/service that uses your chips and hardware rather than finance rackspace. I feel we are seeing the start of this now.
actually lunar lake only catch up to i3 level performance but it cost very high for laptop manufacturer, due to memory on package, TSMC N3E tech, difference package.
Microcenter is insane.
27:09 most people I know who want to go this route, with a " starter APU" aren't truly trying to save money. What they're trying to do is extend the purchasing timeline while still having a perfectly workable system that at least let's them enjoy it. So for example they really only have a $1,000 to spend today, but will have another 1,000 in a month or two. Or if they are trying to save money, it's because they're trying to buy piecemeal on sale certain items. Like the case, power supply and motherboard/memory at one point in time, knowing there will be probably sales or a new version or better version of the CPU later, and the same with the GPU. Because it's only a few months before the new release comes out but they would like to have something for the summer... Envisioning their actual end state being in about 6 months
I know the packaging will definitely be more expensive but will Lunar Lake really be expensive to produce considering how small it is?
AMD has had deals with ASUS since the white box Athlon 500\550 bundle... They had the first AMD Advantage laptops (and mostly only)... That's why Strix Halo is a necessary product... 4070 perf in a single chip solution...? They can cut Nvidia out of more laptops since why put a 4060\5060 in when you can save on cooling and complexity...
I can only assume that the usual suspects will make killer MiniPCs with it next year...
49:02 We said something positive about Intel
Literally the story of this channel 😁
Yeah the hate is deep.
Go buy the house burning i9-14900K, and then don't cry to me when you need to RMA it. I'm not gonna make bad recommendations just to be "fair".
And clearly, you have no memory of this channel...which is basically the only one that leaked dozens of Intel products and was hyped for ARC until it was clear it would fail (and it did btw).
@@MooresLawIsDead Lol, see the reply from this 'unbiased' channel. Don't steep this low.
Really want strix halo, but might buy strix point
It'll be such a premium but the upgradable ram can be a game changer depending on the usecase. Mid range GPU with unlimited vram ❤❤❤
"Secret covid spreader"
Now that's funny xD
I just went out like nothing had changed. Got it 4 times, emergency room once and laid out for two weeks doing nothing the second.
You saw me out in public and you knew I had it LOL
Never changed my habits and never had it. Plenty of people in my life had it.. my immune system is usually better than most though.
@@christophermullins7163 My wife's seemingly immune.
@@rainbowbunchie8237 I am in a different customers home every day so I am subjected to all manner of germs daily. Been doing it for 20 years. I'm sure she is happy about it.
I could def see Zen 5's being almost half the power at same performance, Zen 4 is using N5, Zen 5 is using N4X, which is the 5th iteration of TSMC's 5nm process, it should be much better
@@benpedi6645 did u read my comment properly, i'm saying that Zen 5 should consume almost half the power for same performance
(Edit: in tom's scenario, altou 7800x3d almost never runs at it's 120w tdp)
Any information on how the 5800xt and 5900xt might perform? I'm especially wondering if the 5800xt at $250 would perform significantly better than a 5700x3d which retails for $200 new right now since it boosts higher but doesn't have 3d v cache. Also any idea if there will be a 5800xt 3d?
The metaverse was never a thing, no one wanted it to be a thing. VR and metaverse are not the same.
I'm still surprised that AMD hasn't used their increased profits and engineers to help UE5 add FSR to the engine so that it's available as an option to developers. Even more surprised that Nividia did do that a long time ago. Maybe now that AMD is making it be in separate coding libraries, they will work with UE. And I'm talking about a built-in option not a requirement, still leave it up to the actual game developers.
If Strix had full Zen 5 AVX-512 it would be the perfect emulation APU
Should i wait for Ryzen AI HX 370 based G14 or G16 or go ahead and get the 8945HS Oled G14, is this AI NPU power helpful in the long run? It already gas a 4060 or 4070, so doesnt matter if its 780m or 890m
AMD's choice is to use their TSM advanced packaging allocation for $400 consumer chips vs $20,000 AI chips. Let me guess how they will use that allocation ...
The Problem with the ZEN5 Efficiency thing is that a 7700X loses 2%, yes two percent performance going from 105W TDP (142W PPT) to 65W TDP (88W PPT) so when somebody from AMD says that ZEN5 8 Core doesn't need 105W TDP, that says nothing about it's Efficiency. ZEN4 X SKUs were just pushed without reason.
Pushed without reason? Competition is reason?
@@Games_and_Tech Nah, for 2 % it's just silly, also the direct Competitor (13600K) was/is still faster in MT. For 7950X it kinda made sense to use 230W instead of 142W, because it gave 5% performance and that made it slightly beat 13900K in MT. Still silly, but not absolutely pointless like 7700X.
@@phil.s.3362 they are competing against intel, who is going for almost 400w... then it make some sense
First thing i do when i get a new CPU is check the headroom for fun and find that there is not much as there at there limit, then i pull back the voltage to find the sweet spot for perf and thermals, and im good to go, 2% ipc may not even be a fram or two.
Is Dan ok? 😂 for old times😂
Yeah, X elite needs to be $1200 at the most. Most models need to be $699-999 AND have 32GB of RAM.
i mean i understand why they didnt disclose efficiency on the strix halo.... theres nothing to compare it to. the 8000G apus are gonna be blown away in the graphics department by the strix. i would call it efficient in any context because you dont need a gpu for low end gaming
The killer feature of these APUs, compared to the same expenditure in CPU + cheap old GPU, is power efficiency. At European electricity prices, an extra 100W in 8 hour/day use can burn $50-$100 in a year. Add that up over a few years, and the APU makes a LOT more sense than the CPU + old GPU option.
Also, cost aside, burning electricity 'round here is still contributing to Putin's war chest, and of course it isn't much help to the planet either.
I'm rocking a $200 8600g for now and it's just fine.
When speaking of strix desktop I'm assuming this embedded small form factor systems with no graphics slot? Not a socket.
No canned fish - got it!
Sending canned muscles!
Quest 3 is REALLY good - I think the problem is Quest 2 is just “ok” so it can turn a lot of people off.
The 7800x3d can pull 90w max. Buy if it’s around 65w that’s still really good
At 1hr into the video, during the X3d discussion it's mentioned that X3d is launching at the same time as non X3d.. Wait, what?!?!? I'm pretty sure that this is not the case.
The price range shown is simple, the Left column is 9000 series, the right column is the price of 9000 X3D series. lol.
9:48
Super minor correction for how to say “Weibo” if you care: it’s pronounced “way-bwo” :P back to the video for me
How much work would it take to get a medusa Z2 and medusa console chip😮
It is four weeks (previously two) to Strix Point launching. Asus has three models announced (S 16, ProArt 16, G16). There's some Geekbench results for an unnamed Lenovo model. Acer has a Swift 14 that also just showed up in tests. That's not a lot of models, even compared to X Elite. Moreover there are just two APUs to choose from, again even less than X Elite. This doesn't really feel like a launch with massive volume.
True. However, it's important to point out several skus is good for OEM and manufacturers, but not in anyway desirable for consumers. Two is plenty. Six differents X chip is a drag.
@@mickaelsflow6774 Lots of people would love to see 8-core Ryzen 7 AI's and 6-core Ryzen 5 AI's.
The fact that so many people are foretelling of doom because AMD looks to only be releasing a 70 class card this gen really shows how powerful the halo effect is in the GPU market. It's like thinking every car manufacturer is giving up and going home because they aren't competing with Bugatti's latest hypercar. Just because something doesn't compete at the absolute top of the market doesn't mean it isn't a fantastic product in its own right.
@@djchristian82 not really. The architecture work for rdna4 was already done. The real reason it was cancelled is that consumer graphics has the lowest margins in entire amd and the very complex multi chip module high end rdna 4 gpu woud have competed with limited packaging capacity with high margin data center products.
I hope lunar lake with 16gb ram comes in thinkpads with 90wh batteries. It'd be great as a mechanic diagnostic and module programming laptop.
3500 Timespy for a 15w 890M GPU is the same as a GTX970, min spec for older VR headsets, Valve Index, Reverb G2...
I'm starting to notice prices dip a bit. I saw an INTEL 13700 with NVDIA 3060 GPU (the ones with 12 gigs VRAM) for under $900. I was tempted to buy it, but I stood my ground and will wait a few more months. I still think prices are going to come down a bit more 🙂
I don't need a top of the line computer. I just want one that will be good for 3 or 4 years before I buy another. Usually you can do that with an $800 to $1000 computer when they are on sale.
When is it likely that A desktop APU reaches 3050 performance?
Is it likely to ever get an x3d APU?
I can corroborate the claim of late supply for X elite models - down here in Singapore we only received snapdragon laptops *this week* for retail shelves. Just Asus S15 models
I'm not a gamer but I watch this 'cause I love tech, and, I'm watching this in VR lol.
I work at home in a cramped little room, and I much prefer working in a Virtual Desktop environment that supports dual monitors now and where I can switch seats and watch TH-cam on the big screen in my fancy VR apartment where I can look out the big windows and see the cars and people moving about in the streets below.
I also like Immersed and I'm interested in their Visor, but Virtual Desktop is less friction for when I want to put on my Quest Pro and use my treadmill as a standing desk or for slow walking when working and fast walking when watching stuff on the big screen. The Quest 3 can support 3 monitors in Virtual Desktop with much better resolution.
It helps me because I can get exercise while working, don't need more monitors for that, and can feel less cramped and can focus better on coding.
If someone calls me on Teams I can just put down the headset (when using Virtual Desktop) and sit down at my monitors. If all my team were to work in Immersed that would be viable too.
I say I'm not a gamer, but I do like Microsoft Flight Simulator in VR which is meant to be better supported in the new release later this year, and I do like particular applications like the kayaking simulator thing. I tend to play those using Steam VR with Virtual Desktop. I use my desktop's WiFi card as a hotspot.
If I had space I would like things like beat-sabre and virtual table tennis etc. but I don't have the space.
Anyway. VR is important for me. Very important! It's what loosens my grip on money lol. Though I do want the RTX 5090 for AI lol. I guess I'm very much a minority. I must be an "alien" like Mark Z. :)
nb: I went from moving from desktop's to mini-pcs / NUCs etc. which were great for Proxmox and all the IT and Dev requirements I had, to a desktop again ... I7 Raptor Lake and RTX 4070, just because of AI as GPT 4 came out last year, and with VR being a plus. I bought into it then. And my enthusiasm has only increased as VR and AI do deliver for me. I realise AI, as per my understanding and experience with it so far, will likely completely replace Devs eventually with a number of job-transitions leading up to that in the shorter to mid-term, but I'm still enthusiastic for that future even with the surety it will delete my job and my main vocational skillsets. I'm passionate about tech - actually working on projects ... hardware and software and IT etc. These innovations provide solutions to the projects I'm working on - enabling and facilitating. I'll just need a new source of income if Universal Income doesn't become a thing. "AI" models and architectures improve all the time ... already I can do more "AI" stuff on my 4070 now than I could a year ago ... and I hope the same trend will continue with the 5090. It will do a lot for me in the short-term, and a lot more for me in the long term.
How can Apple Vision Pro have low sales, when they didn't even started shipping in EU/Germany?
Eu doesn't care about apple
@@Rojfos that sounds rather like apple does not care. It's not EU which prevented them, it's just apple which was neglectant to produce enough to supply the other market which is as big as the US. Or EU was Plan B incase they can't sell enough in the US
Strix looks good for minipc's. I have one with an 5800H and its a tad too slow for 1080p gaming, but Strix should be more than fast enough for that.