• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Share your AIDA 64 cache and memory benchmark here

New best:

VDD: 1.55v VDDQ/IO: 1.30v VSOC: 1.24v
Still working on closing the safe mode gap of 1.9ns via OS optimization.
Goal is 55.xx - Seen others with just a .2ns gap normal v. safe mode.
Aida64 - 56.8ns 6200 cl26.png
Aida64 - 54.9ns 6200 cl26 SAFE MODE.png
CPU-Z 7800X3D.png
PYPrime_2B_7.455.jpg
ZenTimings_Screenshot_6200 cl26.png
 
ZEN 3 in action. Think I should start Ocing ram again its been awhile. lol
cachemem 497ns.png
ZenTimings_Screenshot PV4000 4066 CL15 49ns 2024.jpg
 
New best:
VDD: 1.55v VDDQ/IO: 1.30v VSOC: 1.24v
Aida64 - 56.7ns 6200 cl26.png
ZenTimings_Screenshot_6200 cl26.png
 
My current settings on my X3D:

Capture.PNG


And an old shot of my fully engorged 5600X :D

5600X.JPG


:rockout:
 
Update: Final round tested fix my 4x8 Vipers 3933CL16 GearOFF low voltages Wheas free, thanks again to legendary @Veii for support here
toast-v1.gif
~Never give up~
wut.gif

Screenshot_20240214_093628.png

IMG_20240214_204842.jpg
 
Last edited:
Best overall profile:

Daily Stable.

CPU: BCLK 105 - LLC:6 - Scaler X8 - Boost Clock (-) 175MHz - Time Spy CPU: 15089 Ranked 34th of 104,956 results.

RAM: 6200MT/s CL28 36 32 30 / 2200FCLK
VDD: 1.45v VDDQ/IO: 1.30v VSOC: 1.24v

Aida64 - 57.5ns 6200 cl28 BCLK 105 - LLC6 - Scaler X8 - BC (-) 175MHz.png
Aida64 - 55.7ns 6200 CL28 BCLK 105 - LLC6 - Scaler X8 - BC (-) 175MHz.png
ZenTimings_6200 28cl G.Skill.png
G.SKILL RAM SPECS.png
 
1000002591.jpg


5600x with ripjaws 5 hynix cjr 3200 16,18,18,18,38,56 running at 3600 16,18,19,20,40,60.
 
3800 CL14
cachemem 3800 CL14-14-14 514ns 2024c.jpg
GS 4400 C19 Kit 3 3800 CL14-14-14 TM5 Pass crop 2024c.jpg


4133 CL15
cachemem 489ns 2024f.jpg
ZenTimings_Screenshot 49ns B 2024f.jpg
 
Last edited:
Hi guys, look I found a way to 3933CL15 GearOFF, get better latency and speeds than previous CL16, all fine Wheas free, my goal 55ns
Many thanks to mem Lord OC @The King guess I did it well as you suggested guided me inline
4x8gb 3933mhz CL15 GearOFF 1T
Screenshot_20240218_074929.png

IMG_20240218_080323.png

IMG_20240218_074525.jpg
 
Last edited:
7940x OC/Undervolted by individual core from 4,1Ghz to 4,5Ghz.....Reg ECC memory at 2666Mhz with tweaked timings

7940x_Memory.png
 
IMG_20240221_003513.png


Found this thread after some time and decided to check how much IF my 5600g can handle. After 4800MT/s next step broke windows, although I was only 1.25 SOC . I've seen others pushing SOC to 1.3V but only 4600 was max result. Other thing is that latency improved noticeably only at 4800MT/s ... Freaking windows can't recover after this one ;D this was normal windows11 environment. Not safe mode. Ram kit is g.skill ripjaws 4000cl18 2x16 single rank.

Update: pushing SOC to 1.3V made no difference. Windows crashed every time.

Maybe if i try to loosen the timings or increase voltage. But I wasn't expecting this result anyway.
 
Last edited:
Hi guys, look I found a way to 3933CL15 GearOFF, get better latency and speeds than previous CL16, all fine Wheas free, my goal 55ns
Many thanks to mem Lord OC @The King guess I did it well as you suggested guided me inline
4x8gb 3933mhz CL15 GearOFF 1T
IMG_20240220_213700.png

IMG_20240214_204842.jpg
IMG_20240220_205534.jpg

IMG_20240220_205655.jpg

Hi guys, couldn't resist till next week, Day3 Test #2 happy me Wheas free, thx. again, nice to see near rude very low voltages applied in combo Soc/vddp/ccd/iod/vdimm good to go now:slap:
4x8gb 3933mhz CL15 GearOFF 1T CO-22 All
IMG_20240220_215118.png
 
Last edited:
aida.png


That's pretty good for CL40 rated M-die Kit. Still needs little calibration.

@mrcardio33x

Do you know which exact setting is getting you over 70k read/copy? Its pretty rare to see that even on maxed out 1CCD, most of results are hovering and bumping just below 70 . Is it just super low tcl an trfc or is it something? else?
 
View attachment 335947

That's pretty good for CL40 rated M-die Kit. Still needs little calibration.

@mrcardio33x

Do you know which exact setting is getting you over 70k read/copy? Its pretty rare to see that even on maxed out 1CCD, most of results are hovering and bumping just below 70 . Is it just super low tcl an trfc or is it something? else?
Look at his CPU FSB he overclocked it to 105Mhz that gives you extra performance in AIDA64. Yours looks stock at 100Mhz.
 
Would love to learn how to manipulate/OC memory...

cachemem.png
Zen.jpg
 
Last edited:
Would love to learn how to manipulate/OC memory...

View attachment 335967 View attachment 335970
Unfortunately my experience with DDR5 is Zero, but looks like you loosing performance because of your UCLK clock.
This thread will be a good place to start if you want to learn more about DDR5 overclocking on AM5.

Edited with correct thread.
 
Last edited:
Look at his CPU FSB he overclocked it to 105Mhz that gives you extra performance in AIDA64. Yours looks stock at 100Mhz.

My max boost clock is 5125MHz
CPU
: BCLK 105 - Boost Clock (-) 175 MHz = 5125MHz - LLC 7 - Scaler X6 (Without a boost limit I'd hit 5300 MHZ but would not be stable)

The none X3D chips Boost higher Clocks 5500MHz+ so it's not the BCLK giving the advantage in Read / Copy in my case.
I'm not sure exactly what timings, etc that helps most in this regard, probably a combination of timings.

The non X3D chips hit lower latency due to higher boost clocks.

Adjust tRFC & tFEFI for starters, latency will go down. But obviously there is more to it than just those 2 settings. Running 1:2 is better at 8000MT/S if your ram can do it. Otherwise, stick to 1:1 especially for single CCD chips.
His tREFi is maxed and tRFC isn't going to make much difference, this is DDR5

View attachment 335947

That's pretty good for CL40 rated M-die Kit. Still needs little calibration.

@mrcardio33x

Do you know which exact setting is getting you over 70k read/copy? Its pretty rare to see that even on maxed out 1CCD, most of results are hovering and bumping just below 70 . Is it just super low tcl an trfc or is it something? else?
tRAS: 30
tRC: 67
tRRDS: 4
tRRDL: 8
tFAW: 20
tRTP: 12
 
Last edited:
My max boost clock is 5125MHz
CPU
: BCLK 105 - Boost Clock (-) 175 MHz = 5125MHz - LLC 7 - Scaler X6 (Without a boost limit I'd hit 5300 MHZ but would not be stable)

The none X3D chips Boost higher Clocks 5500MHz+ so it's not the BCLK giving the advantage in Read / Copy in my case.
I'm not sure exactly what timings, etc that helps most in this regard, probably a combination of timings.

The non X3D chips hit lower latency due to higher boost clocks.


His tREFi is maxed and tRFC isn't going to make much difference, this is DDR5


tRAS: 30
tRC: 67
tRRDS: 4
tRRDL: 8
tFAW: 20
tRTP: 12
Actually you are mistaken boost clock only affect latency not read/write/copy.

This is 100% the case on DDR4 so don't see why DDR5 will be any different. @Det0x Maybe you can shed some light on this?

I am sure I am correct. However unless DDR5 works differently set your BCLK to 100 stock
with Zentimings open and see if you can get the same results with read/write/copy. Thanks

Stock FSB 100
fsb 1.jpg


OC FSB 100.75
fsb 4.jpg
 
Last edited:
Actually you are mistaken boost clock only affect latency not read/write/copy.

This is 100% the case on DDR4 so don't see why DDR5 will be any different. @Det0x Maybe you can shed some light on this?

I am sure I am correct. However unless DDR5 works differently set your BCLK to 100 stock
with Zentimings open and see if you can get the same results with read/write/copy. Thanks

What are you talking about?? "Mistaken" about what??

I stated the higher boost clock helps latency not read / write / copy.

I don't you you actually read what I posted..

Seriously try reading what I posted first.
 
Last edited:
Back
Top