Y-CRUNCHER - PI-2.5B Bench

[ir_cow]

quickest way to get a high score is to tune your memory to just above instability. Like it won't pass a long memory test, but a cold boot is fine. Next do the same for the CPU. Cold-boot. One pass > save and post like your a boss Seems pretty common among the top 10 scores on HWBOT.

 

[ShrimpBrime]

quickest way to get a high score is to tune your memory to just above instability. Like it won't pass a long memory test, but a cold boot is fine. Next do the same for the CPU. Cold-boot. One pass > save and post like your a boss Seems pretty common among the top 10 scores on HWBOT.

Oh, not sure if that's just an out-loud comment for everyone, but that was a quad with HT I posted. 38th fastest of all quad core submissions meaning I just killed all the 6ghz 7700Ks out there and I didn't even get to overclock to get it.

Cpu Threads, Cpu & Mem Frequency, Fast Timings, shut off all the background tasks.

Play with Priority will help you determine stability.
If you can run RealTime, pretty stable. If it only passes at normal priority, probably not stable.

I set up my Intel like this for the sake of testing different settings.

 

[ir_cow]

@ShrimpBrime more speaking out loud. Some people will put in the work to really tune their systems. Generally I don't assume most of these top scores are stable. Still fun to play around though.

 

[ShrimpBrime]

@ShrimpBrime more speaking out loud. Some people will put in the work to really tune their systems. Generally I don't assume most of these top scores are stable. Still fun to play around though.

"Benchmarks are drag races for PCs"
Top scores shouldn't be stable. Never expect it.

 

[stahlhart]

"Benchmarks are drag races for PCs"
Top scores shouldn't be stable. Never expect it.

Anything I post that is CPU or memory specific is validated stable with OCCT and the y-cruncher component test, in the best interest of avoiding operating system corruption. I've gotten to a point where I look at benching as more of a stability tool than for high scores, because I don't have the most cores, or a custom loop, and I'll always get outrun by those builds.

I have higher presets in NVCP and MSI Afterburner than the daily driver settings for the GPU for benchmarking that I know will pass everything in 3DMark, but they will generate errors in OCCT 3D Adaptive and will crash something like FFXIV Dawn Trail, but I still use them with 3DMark for the sake of consistency with earlier runs.

Trying these settings out for grins. Second one is the parts build that I posted the Legendary scores for last night.

Looks like I forgot to reset PL1 and PL2 on it after the BIOS update, whoops. 1B for this one because it only has 16Gb of DDR4.

 

[ShrimpBrime]

Anything I post that is CPU or memory specific is validated stable with OCCT and the y-cruncher component test, in the best interest of avoiding operating system corruption. I've gotten to a point where I look at benching as more of a stability tool than for high scores, because I don't have the most cores, or a custom loop, and I'll always get outrun by those builds.

I have higher presets in NVCP and MSI Afterburner than the daily driver settings for the GPU for benchmarking that I know will pass everything in 3DMark, but they will generate errors in OCCT 3D Adaptive and will crash something like FFXIV Dawn Trail, but I still use them with 3DMark for the sake of consistency with earlier runs.

View attachment 360556


View attachment 360557

Trying these settings out for grins. Second one is the parts build that I posted the Legendary scores for last night.

Looks like I forgot to reset PL1 and PL2 on it after the BIOS update, whoops. 1B for this one because it only has 16Gb of DDR4.

Looks good to me. Faster than mine, but was stuck on a B660 board for a while. 1m21s at 5.2ghz memory was at 5600mhz OCd from XMP 5200mhz. Was a couple years ago though. Talking about 13700k

12600K, looks like was able to hit 1m37s 1b run. 4.6ghz no sorry that was 2.5b run. 1b was 34s.

 

[AVATARAT]

@ShrimpBrime more speaking out loud. Some people will put in the work to really tune their systems. Generally I don't assume most of these top scores are stable. Still fun to play around though.

That's in depending, most results here are probably stable, it's not an extreme OC place like hwbot.
My results for example are from my 24/7 configuration, I'm not an extreme OC-er, I just tune my system because I want it to work good

 

[Super Firm Tofu]

Nothing like the extremes posted above, but new daily driver, overnight Karhu/core-cycler/game stable.


2.5b

5b

10b

 

[Det0x]

Have updated my 9950X numbers

Managed to wipe all the LN2 scores and get rank #1 in 1B for a short time before one of them uploaded their backup score, so currently i'm down to rank#2 behind bruteforce cold
But this is atleast the highest efficency run when you take core-clocks/performance into consideration

Still #1 for 2.5B, 5B and 10B tho

Hardware used:

Direct die 9950X ES with SMT disabled for 1B (its only faster in 1B)
Watercooled by MO-RA in a big loop
ASUS X6700E GENE @ public 2303 bios
Gskill royal EXPO 8000 @ 8340MT/s CL32-45-39-46 GDM-disable
Clean install Win11 24H2

1B = 10.443s

2.5B = 28.805s

5B = 63.086s

Links to hwbot ranking:

Domdtxdissar`s y-cruncher - Pi-1b score: 10 sec 446 ms with a Ryzen 9 9950X

The Ryzen 9 9950X @ 5653.1MHzscores getScoreFormatted in the y-cruncher - Pi-1b benchmark. Domdtxdissarranks #9 worldwide and #5 in the hardware class. Find out more at HWBOT.

hwbot.org

Domdtxdissar`s y-cruncher - Pi-2.5b score: 28 sec 805 ms with a Ryzen 9 9950X

The Ryzen 9 9950X @ 5477.3MHzscores getScoreFormatted in the y-cruncher - Pi-2.5b benchmark. Domdtxdissarranks #3 worldwide and #1 in the hardware class. Find out more at HWBOT.

hwbot.org

Domdtxdissar`s y-cruncher - Pi-5b score: 1 min 3 sec 86 ms with a Ryzen 9 9950X

The Ryzen 9 9950X @ 5452.6MHzscores getScoreFormatted in the y-cruncher - Pi-5b benchmark. Domdtxdissarranks #2 worldwide and #1 in the hardware class. Find out more at HWBOT.

hwbot.org

And lastly we have my 10B submission, but for this run i used 2x24GB m-die @ 8400MT/s with more lax settings since the benchmark require minimum ~44gigs memory

10B = 144.904s

Domdtxdissar`s y-cruncher - Pi-10b score: 2 min 24 sec 904 ms with a Ryzen 9 9950X

The Ryzen 9 9950X @ 5850MHzscores getScoreFormatted in the y-cruncher - Pi-10b benchmark. Domdtxdissarranks #2 worldwide and #1 in the hardware class. Find out more at HWBOT.

hwbot.org

 

[glnn_23]

9800x3d

 

[Det0x]

My supertuned 9800X3d

1B

2.5B

5B