Testing out the ASUS Intel Baseline Profile
ฝัง
- เผยแพร่เมื่อ 28 เม.ย. 2024
- My Patreon: / buildzoid
Teespring: teespring.com/stores/actually...
Bandcamp: machineforscreams.bandcamp.co...
Twitch: / buildzoid
Twitter: / buildzoid1
Facebook: / actuallyhardcoreovercl...
#intel #z790 #asus #i914900k - วิทยาศาสตร์และเทคโนโลยี
Apparently clear CMOS on this BIOS version sets the CPU to 280A ICCMAX and 253W PL1 and PL2. I just tried load optimized defaults and then cleared CMOS. And the Cinebench R23 score is 36.5K points. Which kinda sucks for a 14900K?
EDIT: I think I made a mistake in this video saying that better power delivery is why it doesn't crash as much as the Gigabyte Z790 Xtreme X board. The Apex Encore defaults to 0.55mOhm AC LL which is higher than gigabyte's 0.4mOhm AC LL. So it's not better power delivery it's just more voltage.
Yeah 13900KS and I got about the same score on mine. I thought that was a lot. Previous bios with just XMP and Disable - enforce all limits set was 39.7K. I am also watercooling with direct die. Truthfully wasn't having performance issues anyway, was just curious, so reverting back to previous bios with the settings I had.
280A * 4 - 1024=96. Your board is setting ICCMAX to 96. I have no idea if this is AMI's, Intel's or Asus' doing.
Interesting. Intel baseline for 14700K also is 280 / 253w. 36.5K definitely is 14700K AIO OC territory..
Before the hotfix and without any undervolt attempts CB23 dropped my cpu's p-core frequency down to 4.6 GHz (worst case) while pulling 253 watts. That was so weird. Now with the uCode 123 it drops to 5.2, while pulling 220 watts. I don't know what Asus did or if they simply didn't give a dang about K-processor behavior on their B660 lineup, but things are going that well, I refund the Gigabyte Z790 I ordered a few days ago. OC with these chips anyway is pointless anyway :D
what is next ? asrock ? MSI ?
So, being able to run the CPU at stock speeds is the new Intel hardcore overclocking...
yep. i mean they prefix all of their claims with "up to", but sending out sp > 110 chips to reviewers, while knowing that most customers are going to receive sp < 90 chips and that they'll never reach these claims, was the real 8===D - move here.
no, if you tweak your own settings many of them do just fine with stock clocks and good volte/watts - and also fine when you go to overclock. I have no idea why the collection of defaults are doing what they are doing - maybe in the past it didnt matter but they have been turning it up slowly over time in various ways and now they shit the bed.
Apparently yes
My biggest problem with this whole debacle is when you run the intel spec limits that they claim should have your 14th gen i7 14700k cpu boost to 5.5 all core and 5.6 single. Yet, instead it drops it all the way down to 4.8-4.9 boost all core and 5.0 single core. I would put the blame on the mobo vendors if it was just because they were running the boards with too much juice to help get better scores but with the intel defaults it doesn't come close to running the advertised specs. So I would love to see what board and cooling setup the intel team used to say these cpu's can pass on to production. Maybe intel needed mobo manufacturers to boost up voltage to be able to achieve the so called up to 5.5/5.6 boost. Second question is why even have a k series or KS if you can't actually overclock it without running delidded or extreme cooling. I am at the point where I am going to revert back to a previous bios just so I can manually set the specs and still get the clock speeds promised.
Intel: Use the baseline profile!
Motherboard makers: And that would be...?
Are you implying the problem is intel not telling them? The people who work at these other companies are not stupid, they can see what different stuff Intel says, they know how to OC, etc. Intel (and AMD all the time, not some of the time) should be very specific and block companies from defaults settings that are crap - but lets be real here, the board makers are not kids who need daddy to tell them what to do - they know what to do, they were not forced to by contract, so they did something else. Bad intel, Very bad OEM's.
@@xBINARYGODx Only Intel has detailed and vast information about what their chips can handle voltage, current, and power wise so it is very much so still on Intel for not mandating their specs. "Recommended" is not strong enough lol, the motherboard vendors are going to play in the sandbox they're given and Intel needs to specify proper rules.
I have asus and strangely they exactly know what's intel baseline profil . You just switch one setting in bios multicore enhancement from auto to disabled and you're exactly in intel spec. So how's that possible without them knowing. th-cam.com/video/HIubZYwBfPc/w-d-xo.htmlsi=Y8yzeVYr8UC-uJwk
With these chip it is not a silicon lottery, it is a silicon lootbox
You could get a 14900K!
(Or you could get a 14900K that should have been down binned to a 14600K)
We know how low Intel will sink just to scrape some extra benchmark numbers whike doing permanent damage to the CPU.
Asus by default enables tvb voltage optimizations, which is why idle / single core voltage is lower. Don’t think gigabyte enables it
Going through Intels public documentation it seems that:
-Gigabyte is running the Intel Base Baseline spec (125/188 for 8+16)
-Asus is running the Intel Base Performance spec (253/253 for 8+16 K SKUs)
So basically both are running Intels official specs, just not the same one.
188 watts choke the cpu very badly
@@troeteimarsch many of the less caring customers who use these chips and want a little time tweaking them won't care about the difference since games for the most part will just be the same either way.
There is no "base baseline spec" that is 125/188 for i9 processors, you might be looking at prerelease QS specs which had a 188w PL2 mode. Neither gigabyte nor Asus are actually running Intel spec in these profiles, but what's actually killing performance is their use of SVID failsafe.
Ahh 2024 where completing a benchmark on baseline settings is 'winning the silicon lottery'. Intel needs to sort some things out ASAP.
The problem is that Intel in their 14900K presentation, if their numbers are correct, they were NOT using baseline profile. The easiest example would be Cinebench 2024 result where 14900K in multicore apparently is 1.06x faster than 7950X and the only way you can get to 1.06X faster is to not limit the CPU to 253W. Recently HUB did this baseline profile benchmark and using Asus baseline, 14900K performance in Cinebench 2024 is practically similar to 7950X and not faster. So basically if I buy an Intel 14900K and I can't get like around 2200 in Cinebench 2024 because either it is not stable at the old default/optimized profile from the mainboard or it degraded that now it is not stable unless using the baseline profile then I will definitely request a replacement otherwise I would say that Intel is lying.
If you run Intel specs you still beat the 7950X, you only tie the 7950X if you enable SVID failsafe like these dogshit "Intel Baseline" profiles do, because you overvolt the hell out of the CPU and hit the power cap quicker.
BTW, you can check the AC/DC Load Line with HWiNFO under Central Processor (not the sensors view).
I thought the AC/DC load line would be under the "Back in Black" menu.
Sorry, I'll see myself out... 😂
and iirc, asus usually give 0.4/1.1 (1.1 seems to lower down the power draw reading a little bit, you can sync it with LLC 4 at 0.98).
For more hardware or ln2/dice
"Clear CMOS (but not really)".
The reason these motherboards have such dramatically different results is due to their VRM used for voltage regulation and delivery. How these motherboards handle voltage is a top secret of each manufacturer, and it is especially critical for new processors as voltage needs to be delivered rapidly and steadily in nanoseconds. The worse boards with inferior components and design simply cannot keep up with the voltage requests from the CPU. The simple solution is to provide a constant higher voltage to disregard the CPU requests.
It should be listed as follows: (or similar)
-Safe boot default
-Intel optimized default (spec to rated voltage, current, power, temp to conservative values, ie. no over voltage)
-Any other boost settings
Between Intel and the motherboard manufacturers redlining the designs and components; heck even over volt and over current out of the box. They are basically doing what we overclockers use to do for years. But at least we use to test out a chip to its max and understand what to look for and fine tune it. Even back down a few settings to get a stable system. Intel and board partners don’t know or give a damn; they just add more voltage and clock higher. Stability, reliability be damn.
If you use asus octool from shamino you can see the ac/dc load line. It sets it to .49/1.1 by default
Thx for the vid and great job as always.
Heard these issues were tied to PL1, PL2, and ICCMax, got some info that this is actually a VRM issue with overshoot when running multi core frequencies. So last night I set my P cores to 5.8 static and turned off boosting while having all the PL settings on unlimited and ran all night on warzone with no issues at all. Will also run cinebench r23
Thank you for making this video!
Another great video with brilliant information thanks for this!
What Intel meant with getting the load-line calibration values as low as possible is not that the manufacturers should just enter low default values into their BIOSes, but rather that they should work to optimize their PCB designs so that the load-line values reported by the Intel tool are lower. The values represent the current-dependent voltage drop caused by the PCB (from both resistance and inductance). The manufacturer has to enter these physical properties of the PCB into the BIOS faithfully in order for the VRM to work right. It's pretty bad if the manufacturer enters a wrong value, both when it's too low and also when it's too high, as that will essentially give random voltages to the CPU which have nothing to do with what the CPU actually requested. Intel unfortunately can't just require the manufacturers to enter a certain load-line calibration value as it's entirely dependent on the actual PCB design. Always forcing the maximum value is also really stupid for the same reasons... It's still wrong, just in the other direction, and the CPU still doesn't get the voltage it requested. (Depending on the implementation, a too-high AC load-line calibration value can also cause voltage undershoots, particularly when the VID is scaled down, so it might still crash...)
It's really rather weird that the Gigabyte board changes its load-line calibration values with that "Intel profile" turned on... Enabling a profile doesn't suddenly make the power planes on the PCB thinner. The values should stay the same.
And it's really strange that AHOC, as a Hardcore OC'er, would keep advocating for NVIDIA style lockdowns; similarly it's really strange hearing him say that MB vendors should all be forced to run very high (near max spec) Load Line Calibration settings (as in the BIOS setting, not the hardware impedance values) *and* lock those settings so that users can't tune them, after all the videos where he's stated that Load Line Vdroop is a very good thing and that users who run high LLC settings are idiots.
Yes the MB vendors should obviously get punished if they're not following the secret design docs to a T, and Intel should really get spanked if they didn't validate their specs or if they sold CPUs that don't meet those minimum specs.
I also really dislike this mass media push for more restrictive power limits and BIOS settings, because I *really* don't want to go back to the years of sub 200W CPUs where you can't get any substantial performance improvements from direct-die cooling with -20C liquid. So you just buy whatever "Extreme Edition" CPU and are out of luck for the next 3-5 years if you need more CPU performance in a game (to avoid literally throwing up from the frame times jittering between 6~9ms).
I admit that I'm biased, since I'm one of those (stupid, yes) people who "upgraded" from an SP 98 13900KS to a 14900KS (turned out to have an SP = 109), which I de lidded, sanded down ~ 200um off the die's backside and ~ 400um off the waterblock, then finally had a shop lap both flat and smooth to < 500nm and < 100nm RMS respectively.
But still, having access to all of these settings really made a tangible improvement to my system's performance in DCS, with the cooling required to keep 6.4GHz @ 1.6V below 40°C at "idle" and 50°C in gaming loads (HT, E-Cores and all power saving settings are disabled. Windows power profile edited in BIOS to ensure the CPU keeps at least 2 cores fully active at a constant 6.4GHz at idle and then verified over 30min at idle, with network cable, monitor and mouse unplugged, that the CPU keeps pulling 200~220W with Vcore measured by multimeter to stay between 1.55~1.58V)
You don't sound very smart on this. You realize that the manufacturers get handed a fully complete schematic from Intel, and it's the manufacturer that is tweaking the design to try to make it unique. It would make sense then, that there are specs for load line calibration, how voltages/current are monitored/tuned by the motherboard.
@@CookieManCookies It's not the schematics that determine a board's VCore power plane impedance. The schematic only determines *which* things are connected on the PCB, but not *how* they're connected (trace thickness, trace length, number of vias, copper plating weight, etc). The physical construction of the board is determined during the PCB layouting phase, which means that two boards with identical schematics but slightly different physical component placement can have wildly different load-line calibration values.
If you look at boards from different vendors, you'll see immediately that the PCB layouts are drastically different even as you go up or down a single vendor's lineup. Heck, even the number of PCB layers varies. As a result, the power plane impedance is also never the same.
its crazy how much you can customize settings on intel chips though, i dont have a good 14900K sample by any means but locking it to 240w (on aio) and tuning ac/dc and vf curves gives me 40.8k score in r23 with 67 degrees on my lf2.
I always found that running intel chips stock, as in just throw them in, is kind of a waste.
HWinfo64 gives AC/DC values under processor.
(when both checkboxes unchecked at startup. most people overlook these)
The clear CMOS vs Load Optimized defaults behavior might be due to AMI APTIO has separate "Failsafe" (after CLR_CMOS) and "Optimized" default values.
Thanks for the info !
What's the difference to the "Enforce all Limits" profile which has always been there? Wasn't that basically the baseline, what is the point?
i wondered the same thing
14700k SP76, r15 runs with no issues optimized defaults on Asus z790itx board. Once I started with cpu occ, r15 was main indicator of system stability because every other stability test was passing (p56/e44/r49/ram 7600mhz @ 1.28v under load) - occt every test config, prime, real bench, y-cruncher… but I had directx errors and occasional bsod, so voltage had to go up to 1.32 to get r15 to run with no issues and eliminate dx and bsod problems.
Same no instability with PL1 250w PL2 250w on 14700k. Runs 1.3v. On a z790 gigabyte.
With a sp rating like that on a 14700k there’s absolutely no way intel are binning these chips properly.
@@exception6651 my 14700k is SP68! I'll never buy Intel again.
@@MrOranj ditto. X341M-Batch here.
I agree it doesn't seem like a degradation issue. 12th Gen are also affected & it shows in a different way: LAGG/Delay through all CPU processes - easily confused for Ram/Ethernet lagg until using certain software & CPU processes (Game Engine) / Older version of firefox. 12900KF I noticed it was at 60% CPU load but also 54% (Grain of salt windows measuring because "didn't care" at the times). The Asus Z690 3401 bios helped. 3501 bios with "intel safe" actually seems to perform the same. Prior the issues happened. So either 1 issue or 2 in what I've seen on 12900KF & 13900K because while 1 freezes out & the other crashes out it happens in the same processes & when it happens it really reminds me of dealing with memory leaks.... but that's my opinion from my experience.
A MUST-SEE for every i9 13-14th gen/Z790 user! Thank you very much for your comprehensive testing! Appreciate your input very much! Buildzoid, do you happen to know what the default Asus Apex Z790 (white-one) LLC behavior is on Auto SVID? Maybe you have mentioned this before, but still. I'm quite confused about setting my setup correctly.
When tuning settings for a 14900ks, or really any intel CPU, is it better to start with defaults, or load intel baseline and then tune from there? I'm not sure on exactly what ALL the baseline profile changes.
Could you consider making a video on bios setup for 14900/14900ks. You made a great overview few years ago on ASUS bios settings. If you could make a shorter one focused on CPU and current ASUS bios would be great :)
That would be nice
What happens when AC loadline is lower than DC loadline? should they always be equal or is it best to leave DC loadline at default values and just lower or raise AC loadline as needed?
Do you have any idea if this is an issue on EVGA Z790s? We just got our 14th gen bios finally
the best thing to do from my experience is set a current limit to 307a 253w pl1 and 2 and set svid behavior typical scenario and set max turbo to 57 for my cpu i am also -0.05 vf point at 5.7ghz and it runs super cool with 1.32v ingames
I guess my 13700k with sp94 is actually pretty decent then.
it makes sense to boot with intel baseline.. as most users wouldnt bother with tinkering in the bios, so it gurantees a "default" out of the box setting..which makes a lot of sense as previously the setting are not so "default"
I remember settings being "sticky" and not cleared on Gigabyte B450 within the AMD settings (iirc) as well. Does anyone know how this is done on a technical side? Is the information being written to a unused section of the eeprom? Or does it overwrite an already existing part?
I believe it's only applying differences from one profile to the other (8x32MB = 256MB UEFI). But when some of the changes make other settings obsolete (like negative offset paired with wrong microcode on Intel) it simply gets ignored. The bios will show that it has been set, but software will tell you that it hasn't.
@@troeteimarsch
but profiles aren't persistent through CMOS clears?
@@tbbhgs every profile but the active one, at least on my board
@@troeteimarsch
Which board is that? I had various board from ASUS, Gigabyte and MSI but usually CMOS clear = all profiles gone.
@@tbbhgs b660 gaming g. Maybe I'm wrong
It would probably not crash with Icc stock current limit since it would throttle way earlier and avoid the huge temperature hotspot overshoot. I like getting focus on this problem but Icc limit is there for a reason. You can have internally temp overshoot because there is only a limit number of sensors inside the chip, so having a the current limit on can help avoiding a guaranteed temperature overshoot.
intel baseline profile is probably not a safe method either.
This is because a case recently emerged in Korea where the 14900k with the msi boxed cooler (pl1 pl2 253w, cpu current limit 307A) option broken from October 2023.
I think turbo boost 3.0 is the cause. I think this momentarily increases the clock speed of one core to 6.0ghz and causes damage by applying high voltage at the same time.
OOB on the Apex Encore is LLC 3 with a AC_LL of 0.5 and a DC_LL of 1.1.
At least it used to be before the new profile.
I thought my 14900k was bad... lol.
Hey i have an issue regarding my pc my board is 640f asus wifi one and my cpu is 7800x3d , when i enable expo and leave from bios it works fine and run games etc but in 2/3 days pc wont boot and gives me orange color on board, i tried finding a solution but i never found any and i came cross ur channel, also the only thing worked for me is enabling expo and putting the profile “expo tweaked” and i tried expo ii/i they all have issues just bot tweaked one and i did the ram benchmark and gave me 76ns on test latency, also forgot to mention i tried disabling memory context restore and it worked but the ram latency is still at 76ns idk if its good or bad i just wanna achieve my ram speed and hopefully u help me with this, my rams is g skill 16x2 ddr5 6000mhz and in bios says 4800 hopefully u can help me or some expert i am willing to reply any information needed in the replies thanks!
How do you unstick or remove the Intel baseline profile?
What's "sp" and how do I check what sp is my 13700k?
Planning to get APEX Encore and 14900KS along with 8400MHZ G skill , is it safe to leave ASUS Multicore Enhancement ON but change the Power limits manually? this will provide me better performance than turning ASUS MCE OFF?
I want to understand not every 13th and 14th gens intel i9 Core have this instabillity issues? only lowest bins? like if I were lucky and win the silicon lottery and get around the highest SP score available I would not encounter instabillity issues? my previous 14900K was SP100 and got the out of video memory issues and BSOD sometimes but I could not run The Finals , every time the game was crashes btw I sold it and planning to try my luck again....
manually oc/under clock and set static voltage with sync all core. Also, 8400 is not gonna work unless you get EXTREMELY lucky or if you can get hold of an A04 white apex and get lucky with the IMC
@@Boogeymanjw I talked to a friend which were get around 107 SP for 14900KS and run 8400MHZ stable with apex encore, why turn sync all cores will disable the Turbo boost of 6.2GHZ right?
It unlikely you'll be able to run 8400Mhz, 7800-8000Mhz is the most likely outcome.
please don't use static voltage unless you are never hitting Tcc max. temps are up, multipliers are down and voltage rises - happy thermal runaway!
@@troeteimarsch I didn’t understand why use sync all cores?
Wonder if Manufacturers could have done some of this tuning based on Intel CPU's of various quality/condition levels, with Gigabyte ending up tuned for CPU's that perhaps should be considered dead instead?(or vice versa, very good CPU's?)
So basically some people got shitty i9 examples that can’t run what the motherboard manufacturers want it to run
*What Intel wants it to run.
Intel was intentionally vague and encouraged mobo makers to push out of box performance as much as possible. This is a failure on both fronts.
Basically Intel pushed these chips too close to what they can reliably do at scale to compete with AMD and we're seeing the ramifications of that. Intel's baseline operating voltages are significantly higher than ever because that's only way they could pull it off and in some cases even that is barely cutting it.
I guarantee you there are some frustrated engineers at Intel right now who knew this would be a shit show but nobody wanted to listen .
I don't understand CPU core/cache current limit Max. 307 my i9 13900k 307 is right? and HWINFO64 where is 307??
yours is within the top 5 worse 14900k SP (91) that i have seen. Beyond that, we are getting to the same conclusion, the voltage regulation plays a (pivotal) role , it is not even the pl1/pl2 or degradation (Assuming it exists) rather if the motherboard voltage regulator reacts to spikes. So a "worse" mb, or to be more precise, one with a worse voltage regulator will end up causing instability, and by limiting icc max that spike is contained within a value that the board can handle.
so you need either a rly good silicon quality or a rly crappy cooler to stop it from boosting too high
My 14900K has a general SP of 106, 115 for the P cores, and 88 for the E cores. Could it be considered as a KS? I know the P cores aren't that hight but the E cores seem to be impressive, at least from my perspective.
Sp106 is a very nice chip.
Mine SP is 96 is it possible to increase it?
@@gvido2923 No. Sp rating is based off the SVID table that intel fused into the chip at the factory.
At 37:59 the Vcore sensor reads 0.186V!
What, How, Why?
if you have an all core overclock will lowering the amps-current effect instability of the already stable overclock?
intel baseline is a problem for me on 13700k. my cpu require more voltage than it gives.. sp is 93 on p cores and 61 on e cores. together its 82 only my cpu require 1.359 voltage but intel fail safe oh damn it drops like 1.3 and under....
Apparently the bios v. 3501 for the Rog Maximus Z690 Formula has a bug that makes the whole BIOS UI lag/stutter
Constantly 1sec. stutter between every second
Managed to downgrade the bios to 3401 and stutter is gone
Going forward intel needs to come out with a single spec. This CPU is 253W. Exceeding this power limit is overclocking. Motherboard manufacturers may configure lower power limits but MUST state the power limits and any applicable TAU durations if they cannot reach the spec. Otherwise WTF are we even buying?
"Here's 8 p cores, 16 e cores and who knows how it will perform and for how long it will continue to do so."
I think the reason why different vendors have different baseline profiles is that Intel changed their profile between 2021 and 2023 multiple times Igor has an article that shows 3 different profiles during this time period.
More videos with your potato 14900k please :)
no
you may have figured out what is really going on while you were musing about it at around 47:10
maybe intel has trouble with instability - so they purposely continue to be ambiguous about the exact specs in order for there to remain plausibility that the problem is something else and not their inability to create stable products ...
My 14900k is below average at SP 95, P core 105, e core 78, mc 80. My CPU can run cinebench at out of the box settings. I’ve seen 14900k with 111 Sp and 125 p core SP.
My 14900k had to be downed to 125 PL1 and 253 PL2 in order to not crash in Helldivers 2. I can’t get above 31k Cinebench 23 score now, and I feel like I wasted $600 on a high end processor. I’m considering returning it for an AMD chip, because this is ridiculous.
Thank you for uploading this video! i don't know if you saw my comment asking for it on the gigabyte profile video or not, but i'ts very useful to see how other boards react. My z-690-e gaming was updated yesterday. Since i updated it i enabled the intel baseline profile and it seems to be running fine. I really didn't have any stability problems before the update tbh (prolly cause im using a 13700k), I did experience some very random crashes in some games like 7 days to die or cs2 from time to time. Most of the time the games will just close without a crash.
Voltages have been pretty reasonable. The highest i saw was 1.46 while playing cs2, and it jumps from 1.1 to 1.3 while browsing or normal use. In idle it gets quite low, up to 0.7v.
Well, Intel just had to beat AMD. But I’d argue the cost to their reputation was by far not worth it, and will cost them far more in the long run, that a few million extra dollars to the engineers to actually engineer a better CPU which is stable at stock settings!
To be realistic it is a miracle that Intel got that far with their 10nm process. TSMC is at 5nm (!) on the Ryzen 7000 series.
thats the thing though they didnt beat AMD even with these unlimited profiles, it was only in some areas but when your still pulling over 100w more than AMDs top offering and can still go toe to toe on synthetic workloads/workstation tasks intel never won imo
@@123TheCloopsure, but Intel is fighting against TSMC that is fueled by endless cash flow from Apple. TSMC is the real competitor, not fab-less AMD. Another question is: Will x86 be the future or will the PC market be dominated by much more efficient ARM designs ? In any case, Intel with its new fabs is positioned not too bad for any scenario. Time will tell.
@@disco.volante If Intel pulled its thumb out of its ass, Intel would be the ones Apple is paying, not TSMC.
@@disco.volante Intel 7 (formerly "10nm") is more dense than TSMC N7 ("7 nm"). The rename had merit. 5nm is only 1 node ahead.
My bios has f10 bug, don't save nothing, just freeze. Help. Mobo Asus ROG strix x670e e
I am on the factory bios on a z790 dark hero and just set the pl limits to 253 and i have 0 problems
Have MCE Off to be safe on 12900KF 4090, likely just not upgrading with all this. Have Asus TUF Gaming Wifi Z690. Not really sure anyone should have to invest any more into LGA1700. Maybe more BIOS updates will solve over time issues.
Please do The Asrock Z790 Nova next. Everyone has videos using the main big 3 brands like ASUS, MSI, GIGABYTE. I would really appreciate a Asrock test.
System date: 12/6/2044?
waitin for msi motherboard!
i had an asus b450 in which the load default or pressing cmos reset wouldn't REALLY reset everything. pulling the cr2032 for a minute is what did the trick
were you modifying the AMD part of the bios or the manufactures part? clearing cmos button only effected the manufacture side, not the AMD side.
So now it's time to test released yeasterday ASRock Baseline Profile :>
so, is 12900k or 13700k 'safer'?
12900K for sure1
haha the bios date
I guess I need some serious help with my 14900KS. Running it on an Asus Maximus Formula. I had it stable in CB23 with an undervolt of -0.005v, all PCores at 5.9, Enforce All limits and using the Intel Extreme power limits 420Amps, 320watts short/long. XMP1 at 6000 MT/S, LLC on 7. (This was all before the Intel stability scandal) was stable in games and benches and would never thermal throttle. All of a sudden, 2 days ago, it would crash in all games (BF2042, COD, Star Citizen, and MSFS2020) Updated to latest BIOS with the Intel baselines and although stable, it runs super hot and benches atrociously at 32k in CB23. It will never boost higher than 5.4Ghz. I have a 360 AiO. SP is 107, with PCores at SP115, ECores SP 77, IMC at 95.
I clearly don't know what some of these settings do, as I'm not an overclocker. Tried to follow some best practices but they're all over the place. I just want to get what I paid for, and am using decent hardware. Also don't want to "degrade" my CPU but I feel it could do better. Any ideas where I should start with the tuning? Should I be looking at the LLC, or any manual settings away from the auto settings? Should I even bother to run XMP for my 6K MT/s ram? Are you running XMP at all?
don't buy intel top end if you're not an overclocker. point blank.
Check your voltages to see if you are running insane voltages.
@@Noah_Aizenthanks yes it was peaking at 1.53 volts at times. I’ve gone and applied the per-core strategy in the video and now runs super cool, boosts to 6.2 on 4 cores, and is stable in what was previously unstable apps (games) However CB23 benches at around 22k which is weird. At work now and typing this on a phone so don’t have any settings but this config is better than I had before. I don’t really care about benchmarks, just That they seem uncharacteristically low.
@@puffman06I buy latest and greatest everything when I can. I keep my PC’s for more than 7 years. I don’t buy a new CPU every 3 months to drag race them for “street cred” on benchmarks. Reason i buy high end CPU is so I won’t bottleneck my GPU’s for a couple years at least. I’m willing to accept that these 14 gen chips run at the ragged edge, and I’ll put in the work to learn how to run it efficiently.
@@shootloadrepeat or you can just buy amd that gives you top end performance and also runs below 1.35v ensuring no voltage degradation ever. 'buying the latest and greatest everything' while not understanding that intel hasn't produced anything great in about 6 years is just sad.
So glad i went with the 7800x3d for my first pc 👍🏻
But it's only an 8 core chip, unless you were looking for the fastest 8 core chip.
You definitely made the right choice! AMD is the clear choice this generation.
@@saricubra2867 If you want the fastest 8 core chip with intel you need to buy a 900K even if you don't need most of the cpu, and there are no more CPUs being released for Intel on their platform. Next gen ryzen will give am5 the fastest everything, in every situation.
@@punishthemeatpocket Raptor Lake Refresh is obviously an epic fail, wasn't needed. It's look like a competition for the worst Intel chip in recent memory, the i9-11900K (i7-11700KS) or the 14900K (the i7-9700K is very close there, it's the only i7 without hyperthreading).
@@saricubra2867intel knew it was never going to be some crazy new thing 😂
Please do an msi vid on this.
0:10 Or you could just take the cmos battery out & disconnect from the mains to clear the settings…
I think Intel xtu should set actual Intel baseline profile or factory certified
Weird that one has to do these kind of advanced things in order to get a darn pc to work properly. And Intel being found out their cpus are dying too early because of these ridiculously over complicated settings.
Dying no, degradation not confirmed. Stability issues yes. Most of social media talk bs to get views and exaggerate to the moon.
I had my 13900k E-Cores degrade on a Asus Z790 Hero.
P-Cores were stable, gaming, benchmarks, where with the E-Cores I couldn't even install a GPU driver, unpack a zip or rar file, Chrome would randomly shut.
If I set the E-Cores to 3.5Ghz stability returned.
Intel RMA'd the cpu after getting me to run all their tests.
I promise you that it's not the motherboards. Even when I limited my 14900k to 125W with all fail-safes and baselines enforced, I still see odd issues. The chip has a major design flaw
look what AMD did to intel, odd that that they can't push nvidia as hard. i love my 6800xt but ....
It will be interesting to see how MSI and Asrock's "Intel Baseline" will compare to Asus and Gigabyte.
Looks like Asus did OK, but Gigabyte's is still a bit dangerous.
Intel needs to come up with a true Base Line. The settings that dictate what can be judged as "stable on manufacturer spec". That should not change between vendors.
Don’t feel bad. I had an sp87. It behaved like it had an sp of 87.
I got SP68. I can't OC it past 5.6.
@@MrOranj I cant get past 5.5 feel me. needs too much vcore.
You guys got me wondering what’s the worst one without being considered broken.
Should there have been a 13850k and 14850k?
Same as you tested 3 cpu. They where all around 87 to 91. After the last bios update from every manufacturer with that Intel base line things, SP bumped to 105 to 107 for 2 CPU
Edit the VID table directly please
Like number 350!!!
Asus sint bad? I've been using from about 3 days now..ita pretty stable
LLC AUTO = LLC3
AC/DC 0.5/1.1mOhm
Svid = intel’s fail change AC to 1.1mOhm
nope, changed mine to 0.4 AC / 1.1 DC LLC3
@@troeteimarsch Watch hwinfo->no sensor-> 14900K -> IA domain loadline on optimized default
Llc3 is 0.47 instead of 0.5 but basically is the same shit
Why not return it?
it's a review sample. It doesn't have a warranty.
@@ActuallyHardcoreOverclocking 99.99% of your viewers would have bought a rattail, so tell them to return lol
i did the baseline but turned mce back to auto and turned svid back to auto ..its stable and running games so far ..i do have an undervolt - adaptive - offset - .1 ...this is fn stupid we even have to do this ! ..r23 score 38,887
Can you let me know if it becomes unstable again in the future? Thanks!
12/6/2044?
Lol i also saw that
Does syncing load line with VRM a valid option now/
13900KS with SP of 112 on performance and 94 on e-cores. weird
Ad OOVM / Shader decompilation UE5 crashes: They are common, like a really common. As I do tech support in my free time I see this basically under every single UE5 game that is not an engine merge from UE4 (i.e. direct UE5 releases).
Under each game there's new batch of users that are reporting this and you can always read "but it's only in THIS game" like under every of these posts. That's dozens of users per game.
Basically any new DX12 game that has heavy shader compilation process or complex shaders will trigger these crashes and you have new batch of users completely unaware that their HW is not set correctly.
Which UE5 games?
here is our baseline profile because we scammed our customers and are trying to get away with it by crippling the performance
SP90 chip here it's so bad lmaooo
I'm RMA'ing, too many issues.
Now MSI :)
You are running multicore enhancement witch is overpowering cpu why??
I just use heat threshold to control the overclock cos im lazy
Holy crap, look at Intel stock, it's on a downward slope from one year ago.
While this is at least something being done, how is your average consumer going to know about this when they buy these products off the shelf? I literally know of streamers that I watch that have 14th gen Intel processors in their setups and regularly see games they are playing crash and they are not tech savvy enough to know of this stuff because an out of the box a motherboard will be set to those "auto" settings by default.
Man my cine score no where near this.
It is sad that Intel has allowed this to happen. How is the default safe state for CPU NOT the default? WTF.
This is whole reason that XMP profile is not the default for RAM. JEDEC safe rating is the standard.
Intel was so eager to allow the vendors to push the chips to getting bigger bars than AMD that they exceeded the margin of error between good and bad CPUs, while shoving way too much power into them.