skulstation

In gpupi whit the tesla m10 no big deviation in speed. just a few ms. What is normal, I have never had the same score 2 or more times with the same HW and settings. Always with a few ms difference Rest of the benchmarks not testet change log is : Features & Improvements - Improved User Interface trying to get a cleaner look - Main window can be resized and maximized - Sorting of the result list is now possible by clicking on the column headers - New Status Bar for loading, warning and error states - Support for docking of child windows (ColdbugMate, Debug Log, Analyze Results) - Windows will now remember the last position, docking state, size and visibility - 7-Zip: - Added Thread Count option to CLI wrapper (thanks to Bastiaan_NL) - Added Timer resolution option to CLI wrapper - Fixed a bug that reported an error in BenchMate when a run was aborted - Any errors in benchmarks will immediately show a Windows notification (make sure you are allowing them) - It's no longer necessary to install BenchMate in the Program Files folder to launch successfully - HWiNFO in 64 bit is now launched in a new unmanaged mode. The app will not be listed in the process list, there are no affinity or process priority settings and it is not guarded by BenchMate at all. This was a necessary step to avoid that BenchMate only works in the Program Files folder. - Improved error messages for users in case the CLI Wrapper can not load the benchmark (benchmark's version not supported, files invalid, ...) - Windows 11 24H2 and Windows Server 2025 are now correctly detected - Added support for AMD Threadripper 9000 series (Shimada Peak) - Added detection for AMD Strix Point, Fire Range, Shimada Peak, Turin and Intel Panther Lake - Added detection for some legacy CPUs like Pentium Pro, Pentium II, Pentium III and several Core architectures - Updated tools: HWiNFO 8.26, GPU-Z 2.65.1, CPU-Z 2.15 Bugfixes - y-cruncher 0.8.5: - hotfix for detecting invalid scores determined by y-cruncher's spot check - shows the progress of the benchmark again (summing percentages) - CINEBENCH R20/R23, GPUPI, all CLI benchmarks: - process errors in run initialization phase will now be properly reported - fixed "Result is invalid" bug due to race condition on benchmark initialization - CINEBENCH 2024: - fixed file integrity validation error with AMD Granite Ridge - Unknown Windows versions will be marked correctly as unknown in the result dialog - Fixed invalid results due to undetected Intel iGPUs - Fixed Bug Reporter, reports will now correctly be sent to bugs.benchmate.org again

Thx mat It works on server 2012

hey @_mat_. I think you forgot to add the necessary certificates for manual installation

links not working in chrome. but i copyed the link to edge and ther they work

A question about no populated second CPU socket allowed The EVGA sr-2 have jumpers to disable the cpu's. Ar we alowed to populatie the second cpu when the other socket is disabled by switching thé Jumper around?

ok i wait , no i don't have a background. i hope thad someone is willing to make on

round 1 starting the 1 june or the weekend after computex the 8 june ?

For me this is the final version. The only thing to pick is a start time and end time. And as suggested by @yee245 and @Leeghoofd it while be divided in memory based rounds Round 1 DDR2/FB-dimm Round 2 DDR3 Round 3 DDR4 Round 4 DDR5

yes

gpupi 3.2

start before or after computex? everyone happy whit the proposed stages? Stage 1: CB20 whit benchmate Stage 2: x265 4K Stage 3: gpupi for cpu 1B whit benchmate and final rules? Max 2 cpu's No remote acces Windows based os only Only workstation / server hardware System pic of open case / open air bench setup All cooling allowed 2x cpu tabs of cpu'z if using two cpu's Standart HWBOT rules apply

FB-dimm whil count as DDR2 As for stage 3 we can use one t'had is not chosen. My preference is: Stage 1: CB20 whit benchmate Stage 2: x265 4K Stage 3: gpupi for cpu 1B whit benchmate

Ok if you also go for memory based stages then so it while be. The use of benchmate was on of the next rules t'had i want to suggest I know that it while be a small amount of users Stages ware a suggestion , i am open for changes special on stage 3 For now this is decided at the moment : Max 2 cpu's No remote acces Windows based os only All cooling allowed 2x cpu tabs of cpu'z if using two cpu's Standart HWBOT rules apply Stage 1: CB20 or 7zip Stage 2: x265 4k or gpupi for cpu 1B

I think t'he most of the modern high end servers in the database , ar from a company and remote login or one time fisikal access Thad why i like to know if there is a way to detect and block remote access ( @Leeghoofd ) Splitting in different ram stages while make this competition a lot bigger than thought, But i let this decision to the moderators.

i think that as long you don't use slot 1 / slot 2 , socket 370 or socket A its possible to run all the proposed benchmarks. In the database t'her are CBr20 scores from socket 604( xeon nocona ) and socket F ( opteron 2*** )

As benchnarks i am thinking of 3 stages ,but i am open for suggestions Stage 1: CB20 or 7zip Stage 2: x265 4k or gpupi for cpu 1B Stage 3: Y-cruncher pi-25b or Pyprime as rules : Any cooling is allowed Standard HWBOT screenshot rules apply Verification screenshot contains 2x CPU-Z tab for the CPU's Add a picture of your benching setup ( no pictutere of a closed case ) windows server allowed Mandatory usage of the Official competition wallpaper ( if anyone can make one ) @Leeghoofd is ther a way to chek for remote desktop login? i like to block remote desktop acces , only fisikal acces

Yes max 2 cpu's to start with. may be next time no limit on cpu's I don;t know if its wise to allow ES cpu's , what does the rest think about ES cpu's ? The normal rules are no ES in competitions as far as i know

I would like to know your opinion. Whether you would like to have a competition whit workstation / server related hardware ? no prices just for fun Just brain storming

Somewhere around 200 and 300 euro max (full box and accesoires) Where are you from? if from the EU, I'm interested

Plz delete this submission https://hwbot.org/submission/5447170_ It's not gpupi 3.3 but gpupi 3.2

merci, à ma connaissance il y en a 3 en Belgique. à moins qu' hybridchiller n'ait vendu le sien. ps google traduction

this score is 1633 not 1363

Rx 7700 xt is A rdna gen 3 Not a rdna gen 1 thad is needed for thé Cc stage

Thx whill trai agen ? When i traid ab it don't do any thing.

what mod's and tool's did you use to overclock the card?