You can pick your game and we can test if you are interested. I can back up everything I say with actual results
Im taking about the 7800x3d ? I dont have one do you ?? you said I have the cpus in question, no need. ???
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
On Intel Raptor Lake, any truth to the rumors that disabling all e-cores hurts single threaded performance of the p-cores??
- Thread starter Wolverine2349
- Start date
Im taking about the 7800x3d ? I dont have one do you ?? you said I have the cpus in question, no need. ???
Soldato
Ah sorry, didn't comment on the 7800x 3d, I specifically said I don't have it so can't test it. Which makes your previous reply kinda weird, doesn't it?
Soldato
I didn't pick one game, i've tried a lot. Warzone / cyberpunk / TLOU / hogwarts / spiderman. TLOU is just the worst case when it comes to power draw, that's why I mentioned that one, my CPU picks at 117w.point still stands you picked one game and almost sounding its the norm ? but I wont get involved you get on with it
could ya get some screenshots up of benches that show this, im interested to know the truth about all this. most reviews just show peak power consumption in rendering or prime etc but not power consumption while gaming or even idle and low load power usage. best to get new thread up and running and we can finally get to the bottom of this cuz at them moment we are just going in circles with no factual data being presented.
Soldato
Well I have this for example, im on a discord call - watching a twitch stream and a youtube video, power draw is between 4 and 6 watts. Usually after full 8 hours of actual work (mostly multiple browsers, excels , words pdfs etc, nothing heavy but lots of apps open) it averages out at around 6 to 8 watts.
During gaming, if you are full CPU bound (meaning 720p ultra with a 4090) highest ive seen is 120 watts in the last of us (the 5800x 3d also draw over 110w, but with much worse performance). Cyberpunk / spiderman, which are also really heavy games are at around 90 to 110 watts. Everything else - even at CPU bound settings is between 40 and 70 watts (hogwarts / warzone 2 etc.). I have multiple videos on my channel showing that if you are interested, I can link you some
Last edited:
Soldato
Sorry I'm on the phone now, I'll send you when I'm home.
Power draw is by hwinfo, I calibrated the motherboard so it's as accurate as it can be (ac DC ll / vid = vcore etc, you can see it from the above screenshot actually. .). The only thing I can't account for fully is vrm loses but those should be minimal since I have a high end board - but they are kinda irrelevant as well if you are comparing between cpus.
I had a 13900k briefly but I swapped back to the 12900k. The 13 part was a power hog in gaming unless you downclocked it. It consumed around 60 to 80% more power than the 12900k. It was fast, too fast, but the power draw was nutty.
Is it the CPU package power sensor under the Core i9 12900K in HWInfo64?
That is only reading I have on mine for power draw.
Soldato
Υes, but my motherboard also has readings directly from the VRMs.
Well I gave up on static all core clocks being something I desire and sold off 13900K and now on a 7800X3D.
So much better and more stable and so much lower power usage and much less heat dumped into case as a result
And 7800X3D has only 8 strong cores and a bunch of L3 cache 96MB and it spanks 13900K in gaming anyways. So much better than dealing with Intel high heat and power consumption and having to disable those useless for gaming e-waste cores.
So much better and more stable and so much lower power usage and much less heat dumped into case as a result
And 7800X3D has only 8 strong cores and a bunch of L3 cache 96MB and it spanks 13900K in gaming anyways. So much better than dealing with Intel high heat and power consumption and having to disable those useless for gaming e-waste cores.
Last edited:
Soldato
What board and memory did you go for with your 7800x3d CPU?
Had you previous experience with a recent AMD build, and did all go well for you when putting it all together ?
Thanks
Went with MSI MPG X670E Carbon
Memory G.Skill FLareX DDR5 6000 CL30 2 X 16GB Hynix M Die with EXPOI enabled
Build went very well and no issues
Did mild manually tuning by enabling EXPO and using Buildzoid RAM tinings:
Easy memory timings for Hynix DDR5 with Ryzen 7000
Text version: https://www.patreon.com/posts/low-effort-rank-77403831My Patreon: https://www.patreon.com/buildzoidTeespring: https://teespring.com/stores/actu...
www.youtube.com
Also set SOC voltage to 1.23 and VDDIO to 1.23 as defualt is 1.3 with my mobo on 6000 EXPO and I read it should be below 1.3 and maybe even below 1.25? SO far perfectly stable no issues.
Though went more conservative on his timings to ensure stability even if all his would work. It lowered my EXPO AIDA64 memory latency form default EXPO 69ns to tweaked timings 62ns.
Also set Curve Optimizer to -20 all cores. I had tried higher like -28 and it passed all tests and was fully stable, but wanted to back off just to ensure unconditioned stability as I heard stability issues can pop up at idle especially with too high negative CO.
And I can say so far perfect stability at idle and under loads at -20. Maybe could go further but not going to push and risk instability as stress tests often do not catch it.
Case in point my reason for switching in particular is that Intel Raptor Lake CPUs when manually tuned despite passing every rough stability/stress test and shader compilation multiple times with no WHEAs nor crashes nor errors made me think I was all set and good to go. Yet a few weeks later doing shader compilation or running CInebench R23 in a game, a random CPU WHEA would show up, but it was intermittent and would not always happen which made me give up on Raptor Lake. Maybe stock settings would have worked fine, but I did not want Intel Raptor Lake for stock and those e-waste cores enabled and also lower clocked P cores. With dynamic clocks anyway, I figured go with the best gaming chip AMD.
It was a shame with Intel as Coffee Lake and prior passing lots of stress/stability tests seemed to guarantee real world usage stability. That sadly was not the case with Raptor Lake as WHEAs were very random indicating maybe easy and fast degradation with Raptor Lake is very real. So sold the chip and mobo and RAM off and now much happier with more reliable and stable 7800X3D even somewhat tweaked, but obviously no static clocks.
Soldato
@Wolverine2349 Thanks for the details, appreciated.
I've been waiting until the RL refresh has been released to see what, if any, difference it would make to pricing.
Both the RL pricing and the AMD, if any, will be interesting to see.
Was it down to cost / performance as to why you went for the 7800x3d over the 7950x3d, including your needs?
Point taken on stability, and how it can be an issue at seemingly random times.
I've been waiting until the RL refresh has been released to see what, if any, difference it would make to pricing.
Both the RL pricing and the AMD, if any, will be interesting to see.
Was it down to cost / performance as to why you went for the 7800x3d over the 7950x3d, including your needs?
Point taken on stability, and how it can be an issue at seemingly random times.
Last edited:
Soldato
- Joined
- 14 Aug 2018
- Posts
- 3,548
You really should have got a 7800X3D from the start as it is perfect for people like yourself that just want to plug and play. I've tested my overclocked/tuned 13700K against my CO/tuned 7800X3D and it was faster than my 7800X3D in all games but one and way faster in general software usage. It's overclocked to 5.8Ghz all core to 6.2Ghz boost and has been for months and I get no WHEA etc.
Raptorlake overclocking really isn't for novices or people that don't have the time to learn/do it properly, much better to get a 7800X3D for gaming instead.
Yes I should have but too late for that now.
Hindsight is 20/20.
I had thought overclocking Raptor Lake to reasonable clocks e-cores off on air would be easy like Coffee Lake was. Turned out not to be the case.
In speaking due you use voltage offsets or static vcore with LLC??
I had always used a static voltage with LLC6. It seems maybe static voltages are harder to get fully stable and require more patience unlike Coffee Lake and prior Intel gen and now you need to learn and have patience to do more testing and use voltage offsets and such? I had it for Coffee Lake 9900K and it was easy and super stable. Raptor LKake is so much harder and I had no patience.
So should have just gone 7800X3D off the bat. I have it now and have not looked back and it is great. No need for superior perf in other workloads. I got just as good of gaming performance as mild tuned Raptor Lake and almost as good as heavily tuned Raptor Lake.
I have been doing lots and lots of testing, quite extreme actually.
I have been tinkering with various hidden power settings that manipulate the cpu scheduler.
Doing tests with cpu-z bench as quick and simple test.
So when we think about single threaded, by default these hybrid cpu's keep all p-cores parked, I dont yet know what determines to get a p-core unparked, but I do know that it will always prioritise the 2 preferred cores, these clock higher than the rest of the cores. If all the p-cores are unparked which can be done by adjusting the scheduler settings to force them to always be unparked (or via software calls likely, as an example all core cinebench unparks them), then single threaded performance will be hurt as it seems to be the parking mechanism that routes single threaded load to the preferred cores, once more are unparked it becomes at the mercy of the standard cpu scheduler.
If the e-cores are disabled in the bios, then some p-cores will always be unparked, and it increases the chance of non preferred p-cores been unparked, I have yet to test with e-cores disabled as personally I dont think thats the best way to use these chips, but I will do at some point, the second likely problem with disabling e-cores is you cant remove background tasks from the p-cores. This I was able to observe a noticeable impact on the single threaded cpuz score.
So using something like process hacker move svchost, browser process, afterburner, hwinfo, discord and other background apps to e-cores via affinity settings. There is also the cpu scheduler master setting which has 5 options.
Automatic
efficient cores
prefer efficient cores
prefer performant cores
performant cores
Now when I tested 'performant cores', it was a lower score vs 'prefer performant cores', likely because when using prefer, it moves lower demanding tasks conflicting on the core to e-cores when a heavy task is running. I hadnt moved every single background binary to e-cores only the biggest one's.
The e-cores also dont just offer more raw grunt, but they also offer more cores to reduce scheduling bottlenecks (can cause stutters in games). This is the primary reason why I think its better to have them available.
The downside of them is if anything interactive runs on them instead of the p-cores then that might give slower interactive performance, and also potentially can be less power efficient, as an example if 'any' p-core is unparked, all e-cores will stay at max clocks due to default power scheduler settings (hidden setting) and this also raises the vcore significantly. So I expect there is no perfect solution to cover all bases, but it is still something I am experimenting with. I have noticed occasionally things run on them when I dont want to, e.g. the UAC prompt always seems to use e-cores to process the prompt box. Unless 'performant cores' forces it over, and I havent found a way to adjust that via process hacker.
Via hidden settings the min number of unparked p-cores is adjustable, so you can force them to stay awake, if using prefer p-cores or forced p-cores these do NOT force cores to unpark, so they only actually work if p-cores are available, luckily keeping both preferred p-cores always unparked has no noticable effect on vcore or power consumption, however keeping 'all' p-cores always unparked does have an effect especially when p-core clocks are increases and as mentioned earlier no longer ensures single threaded stuff goes to a preferred p-core.
If you dont care about the scheduling bottlenecks from having less cores, then an approach could be made by disabling all e-cores, then also setting all cores to same clocks as preferred p-cores (if cpu can handle it, might need a voltage bump), by changing to an all core clock that removes any concerns about needing preferred p-cores for max single threaded performance. But will still lose some for having bg tasks running on them. (my single thread score went down by about 4-5% when not moving task to e-cores)
I have also confirmed as a quick and dirty adjustment, simply using the high performance or ultimate power profile, combined with routing background stuff specifically to e-cores and adjusting the thread scheduler to "prefer performant cores" gives insane performance all over the shop, easily beating my 9900k at everything, but is a bit power inefficient this way when using light load e.g. 30 watts cpu package power to watch youtube.
For what its worth I think its inevitable AMD will bring out a variant of e-cores, on their server chips I think they are just cores with less cache, so might see same on future desktop chips.
I have been tinkering with various hidden power settings that manipulate the cpu scheduler.
Doing tests with cpu-z bench as quick and simple test.
So when we think about single threaded, by default these hybrid cpu's keep all p-cores parked, I dont yet know what determines to get a p-core unparked, but I do know that it will always prioritise the 2 preferred cores, these clock higher than the rest of the cores. If all the p-cores are unparked which can be done by adjusting the scheduler settings to force them to always be unparked (or via software calls likely, as an example all core cinebench unparks them), then single threaded performance will be hurt as it seems to be the parking mechanism that routes single threaded load to the preferred cores, once more are unparked it becomes at the mercy of the standard cpu scheduler.
If the e-cores are disabled in the bios, then some p-cores will always be unparked, and it increases the chance of non preferred p-cores been unparked, I have yet to test with e-cores disabled as personally I dont think thats the best way to use these chips, but I will do at some point, the second likely problem with disabling e-cores is you cant remove background tasks from the p-cores. This I was able to observe a noticeable impact on the single threaded cpuz score.
So using something like process hacker move svchost, browser process, afterburner, hwinfo, discord and other background apps to e-cores via affinity settings. There is also the cpu scheduler master setting which has 5 options.
Automatic
efficient cores
prefer efficient cores
prefer performant cores
performant cores
Now when I tested 'performant cores', it was a lower score vs 'prefer performant cores', likely because when using prefer, it moves lower demanding tasks conflicting on the core to e-cores when a heavy task is running. I hadnt moved every single background binary to e-cores only the biggest one's.
The e-cores also dont just offer more raw grunt, but they also offer more cores to reduce scheduling bottlenecks (can cause stutters in games). This is the primary reason why I think its better to have them available.
The downside of them is if anything interactive runs on them instead of the p-cores then that might give slower interactive performance, and also potentially can be less power efficient, as an example if 'any' p-core is unparked, all e-cores will stay at max clocks due to default power scheduler settings (hidden setting) and this also raises the vcore significantly. So I expect there is no perfect solution to cover all bases, but it is still something I am experimenting with. I have noticed occasionally things run on them when I dont want to, e.g. the UAC prompt always seems to use e-cores to process the prompt box. Unless 'performant cores' forces it over, and I havent found a way to adjust that via process hacker.
Via hidden settings the min number of unparked p-cores is adjustable, so you can force them to stay awake, if using prefer p-cores or forced p-cores these do NOT force cores to unpark, so they only actually work if p-cores are available, luckily keeping both preferred p-cores always unparked has no noticable effect on vcore or power consumption, however keeping 'all' p-cores always unparked does have an effect especially when p-core clocks are increases and as mentioned earlier no longer ensures single threaded stuff goes to a preferred p-core.
If you dont care about the scheduling bottlenecks from having less cores, then an approach could be made by disabling all e-cores, then also setting all cores to same clocks as preferred p-cores (if cpu can handle it, might need a voltage bump), by changing to an all core clock that removes any concerns about needing preferred p-cores for max single threaded performance. But will still lose some for having bg tasks running on them. (my single thread score went down by about 4-5% when not moving task to e-cores)
I have also confirmed as a quick and dirty adjustment, simply using the high performance or ultimate power profile, combined with routing background stuff specifically to e-cores and adjusting the thread scheduler to "prefer performant cores" gives insane performance all over the shop, easily beating my 9900k at everything, but is a bit power inefficient this way when using light load e.g. 30 watts cpu package power to watch youtube.
For what its worth I think its inevitable AMD will bring out a variant of e-cores, on their server chips I think they are just cores with less cache, so might see same on future desktop chips.
Last edited:
Soldato
- Joined
- 14 Aug 2018
- Posts
- 3,548
I use adaptive voltage. I have only ever use constant voltage the very first time when overclocking to establish around what voltage a CPU needs for a certain clock speed.
For 5.8Ghz all core load it will use ~1.394v
Last edited:
I use adaptive voltage. I have only ever use constant voltage the very first time when overclocking to establish around what voltage a CPU needs for a certain clock speed.
For 5.8Ghz all core load it will use ~1.394v
Yikes 1.394V all core workload. No way to cool that on the best air coolers which are the dual tower ones with 120 to 140mm fans like NH-D15 and Dark Rock pro 3 and similar ones.
I imagine you must have insanely good AIO or custom loop water cooling for that.
Neverminded if you also have the 16 e-cores (4 clusters of 4 each which is like the same space of 4 extra P cores as 1 e-core is like same die space as 4 P cores) on which is another 35% power and heat added. You must have insane cooling for that.
I debloat Windows using NTLite. Well do not remove components but disable unnecessary services and such which makes it a pretty lite with regards to background stuff running. I have Windows updates completely shut down so they can not start and thus no high background process usage as Windows updates does take a lot. I also never have Chrome nor any browsers in background open. So no way there scheduling bottlenecks in games. Certainly not in WIN10.
In WIN11 its another matter. I have seen reduced performance with e-cores off even for single thread in WIN11. The scheduler is much more complex and what you describe is likely why. WIN10 never parks any P cores by default. Perhaps your method of disabling e-cores in WIN11 to still get best P core performance works.
In WIN10 you can disable e-0cores set and forget it and no reduced single thread performance. Its the same either way. WIN11 once again works differently due to the way it uses thread director in which case WIN10 does not use thread director.
And yeah I had always used all core fixed clock speed for P cores being all same speed so I should have been covered by last point anyways on at least WIN10 and partly on WIN11 if scheduler was setup right.
Anyways I got tired of Raptor Lake tweaking and went with 7800X3D where I only put a mild -20 CO and tuned the RAM without touching any other CPU settings and could not be happier with its much better stability and much less heat inside the case and just as good or better gaming performance than RL.
And no I doubt AMD ever uses hybrid approach. They have already mentioned they are not. Certainly not on the mainstream to high end desktop CPUs.
What I do want to see AMD do is offer more than 8 good cores on a single CCD. Hopefully Zen 5 provides it, though I am not holding my breathe.
Last edited: