|
|
Post by Aleslammer on Jun 9, 2020 14:45:13 GMT -5
Started this out after reading a bit about Power Reporting Deviation and after playing a little really not sure what is going on as not close to a 100%.? Was just trying to get as close as possible to 100% by lowering vCore actually might not have anything to do with it. Temp -5c fluid on chiller which I was able to maintain at the voltages used. Should be able to see the other information in the SS's. 4.5 CPU @ 1.24v low, 1.272v high, WP1024  4.5 CPU @ 1.264v low, 1.304v high WP1024, CB11.5 & CPUz bench  |
|
|
|
Post by ShrimpBrime on Jun 9, 2020 19:19:24 GMT -5
Oh gosh 0.036% deviation (I guessed number) that no one can see or feel or for the most part give two shits about. On that note, nice job on the chilling and lower voltage!! Minimizing leakage, I like it. My 2700x was 1.188v at 4ghz and was able to maintain sub 20c temps. I think you could get colder with a TEC under the chiller  |
|
|
|
Post by Aleslammer on Jun 10, 2020 12:01:29 GMT -5
Took me awhile to get it through my head how Power Reporting Deviation was being displayed in the examples, Hero I'm using fudging the numbers a little but makes no difference to me.  Thought about using TEC, but if I'm going to go colder might as well go all the way. |
|
|
|
Post by Aleslammer on Jun 11, 2020 13:44:25 GMT -5
This will be my last test on this, 3900x not a happy CPU at low voltage. This is the best I can do with CB20 @ 4.5, in BIOS 1.30v vCore, RAM @ 1.575v, fluid @ -5c. Dew point about 8c today in the house with the AC on.  And yes I do use a little insulation on occasion, doesn't take a lot once you realize there is a hole in the socket area on this board.  |
|
|
|
Post by ShrimpBrime on Jun 11, 2020 21:54:37 GMT -5
Impressive overclock. Good show.
Deviation. Power sustained within margin of error. Lol.
|
|
|
|
Post by sphincter on Jun 12, 2020 20:36:38 GMT -5
Looks like fun! about 8x the score of my old stuff(running warm).  Nice work!  |
|
|
|
Post by Aleslammer on Jun 13, 2020 18:37:10 GMT -5
Doing a little RAM testing with some Patriot DDR4 4400 Viper Steels, not a lot different from the G. Skills I've been using and for the price are a sweet deal. BIOS is all auto except RAM Primaries (as shown in CPUz), vDimm 1.5v, Memory Frequency and FCLK Frequency. You will notice that BM is showing a timer skew, raised BCLK to shown in CPUz from default while in Windows to see where NB died, guess I'm cheating, so the crap score should be ignored.  |
|
|
|
Post by ShrimpBrime on Jun 14, 2020 0:02:39 GMT -5
Doing a little RAM testing with some Patriot DDR4 4400 Viper Steels, not a lot different from the G. Skills I've been using and for the price are a sweet deal. BIOS is all auto except RAM Primaries (as shown in CPUz), vDimm 1.5v, Memory Frequency and FCLK Frequency. You will notice that BM is showing a timer skew, raised BCLK to shown in CPUz from default while in Windows to see where NB died, guess I'm cheating, so the crap score should be ignored. What does the SOC voltage look like? Are you in the yellow yet? |
|
|
|
Post by Aleslammer on Jun 14, 2020 9:00:24 GMT -5
ShrimpBrimeCPU SOC is on auto, looks like 1.08v in BIOS, HWinfo CPU section SOC 1.075-1.087v, MB section SOC 1.072-1.096v, after a couple CB20 runs. Been running 1.15v in BIOS to bench with on this CPU.
|
|
|
|
Post by ShrimpBrime on Jun 14, 2020 9:55:56 GMT -5
ShrimpBrime CPU SOC is on auto, looks like 1.08v in BIOS, HWinfo CPU section SOC 1.075-1.087v, MB section SOC 1.072-1.096v, after a couple CB20 runs. Been running 1.15v in BIOS to bench with on this CPU. Yea, mine on Zen+ doesnt go yellow in the bios till around 1.3-1.35v somewhere in that range. Was just curious if you had it up that high. Im not saying to do it, but since you're chilling, it might not hurt anything. I know the SOC will start pumping heat once you get higher voltage like that. |
|
|
|
Post by Aleslammer on Jun 14, 2020 10:44:01 GMT -5
ShrimpBrime Was trying for DDR4 3733 C12 with a 3600X made Windows a couple times but couldn't do anything, SOC 1.40v vDimm 1.725v. Ended up trashing the OS so haven't tried again.
|
|
|
|
Post by ShrimpBrime on Jun 14, 2020 10:47:48 GMT -5
Aleslammer That's tight timings. A decent accomplishment. I cant get past 3466 on cas 12. Cl14 just seems to play nicer, but Im using a weaker IMC.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 19, 2020 13:10:04 GMT -5
ShrimpBrime Was trying for DDR4 3733 C12 with a 3600X made Windows a couple times but couldn't do anything, SOC 1.40v vDimm 1.725v. Ended up trashing the OS so haven't tried again. I haven't seen any benefit to increasing SOC for faster and or tighter RAM. Have you had a different experience? I also haven't found CL12 to be worth the slightly higher tRFC or voltage. It seems to perform the same in Aida and Geekbench 5 in my brief testing. Attachments:
|
|
|
|
Post by ShrimpBrime on Aug 19, 2020 21:02:25 GMT -5
ShrimpBrime Was trying for DDR4 3733 C12 with a 3600X made Windows a couple times but couldn't do anything, SOC 1.40v vDimm 1.725v. Ended up trashing the OS so haven't tried again. I haven't seen any benefit to increasing SOC for faster and or tighter RAM. Have you had a different experience? I also haven't found CL12 to be worth the slightly higher tRFC or voltage. It seems to perform the same in Aida and Geekbench 5 in my brief testing. Whatcha see here is what you get. Brad has been quiet for a while, so may not get a reply. I'm sorry. |
|
|
|
Post by alpi on Sept 7, 2020 8:41:10 GMT -5
Some low temp testing with my 3800X.
|
|
Nice work! 
