go wild

Compo page at: https://hwbot.org/competition/superpi32LLCintel/stage/5219_superpi_-_32m_with_benchmate

 

5min 7sec 890ms work in progress 

https://hwbot.org/submission/4720358_

Edited April 6, 20214 yr by mllrkllr88

a couple screens from @Audigy's Facebook

so old low mem freq 12-11 ain't bad...

Compo page is up: https://hwbot.org/competition/superpi32LLCintel/stage/5219_superpi_-_32m_with_benchmate,

just let me know with or without BM guys

@Leeghoofd I'd say one combined comp, both with and without benchmate allowed.

There's been some discussion on if Win10 w/ Benchmate is faster than WinXP

Edited April 6, 20214 yr by Sparky's__Adventure

K will add a non BM stage too then

who can help me with efficiency? Learning 32M now... I mean benchmate makes it easy but I'm clearly missing something. I thought its just OS, but I see ppl running Win10 and still way ahead despite less agressive mem.

 

Edited April 14, 20214 yr by CENS

alright I figured it out:

@LeeghoofdWhy do the non-benchmate rules say hwinfo has to be running in the background? Never before have I seen a SuperPi competition mandate this, and it's definitely gonna hurt score.

Also, we already have the Benchmate stage, for the more modern/secure way. Seems a bit redundant.

Edited April 14, 20214 yr by Sparky's__Adventure

This is mine with w10. found that gear1 have better performance . 5:07.487

6 hours ago, Sparky's__Adventure said:

@LeeghoofdWhy do the non-benchmate rules say hwinfo has to be running in the background? Never before have I seen a SuperPi competition mandate this, and it's definitely gonna hurt score.

Also, we already have the Benchmate stage, for the more modern/secure way. Seems a bit redundant.

Do you know or is it a guess? We haven't noticed a 1 sec los with LLC S775. you can disable 98% of the monitoring as we only needed highest core clocks.

Also If its the same for everybody? 

2 hours ago, Leeghoofd said:

Do you know or is it a guess? We haven't noticed a 1 sec los with LLC S775. you can disable 98% of the monitoring as we only needed highest core clocks.

Also If its the same for everybody? 

I didn't know that you could disable parts of hwinfo monitoring, interesting. 

I talked to David and he got a 100ms perf loss versus with and without... you can disable a lot and even slow down the polling

grabbed a board the other day and a 11600k only available here, still dialling in, will install hwinfo soon when I’m fully dialled in. Board working great...
 

27 minutes ago, Bullant said:

grabbed a board the other day and a 11600k only available here, still dialling in, will install hwinfo soon when I’m fully dialled in. Board working great...
 

Awesome result, nice setting!

Edited April 23, 20214 yr by Sparky's__Adventure

On 4/22/2021 at 8:50 PM, Bullant said:

grabbed a board the other day and a 11600k only available here, still dialling in, will install hwinfo soon when I’m fully dialled in. Board working great...
 

 

Welcome back , the OC world missed you !

Joining with slow windows 10 first, it looks like i have weak cpu and weak imc. Efficiency still very bad. Single Rank is like 1.5s slower than dual rank in my test all maxed out at 1:1.

Slow

I think, IMC is OK as most of chips (better than average). Depends on your voltages. I use IO MEM 1.45V and SA voltage 1.5V. Propably I will need more for 3866 MHz tight timings (I got only cycles without right post) With  more weak timings and 2T chip working up to 1900 MHz.

So this is propably almost best what I can do with 3600 MHz (tomorow I will try lower TRFC and RAS) And after I will try 3866 C12 or Windows 7 OS

Edited April 25, 20214 yr by flanker

Got not too bad score, but benchmate said HWiNFO Error, need reboot. Rebooted multiple times still the same, planning to run hwinfo in background after this.

Hello
It's something amazing, isn't it?
Is there any more advantage of Windows XP?
I'm confused
Excuse me because I'm an amateur

slightly better, but are steps by 0.05s to 0.1s :D...

https://hwbot.org/submission/4734729_flanker_superpi___32m_with_benchmate_core_i9_11900k_5min_9sec_776ms?recalculate=true

Why is this result!
I ask for commentary
It's not a criticism

this is propably not in 5GHz challenge (max CPU clock 5600MHz), this is only 5GHz challenge thread

Edited April 26, 20214 yr by flanker

sorry