RapydMark CPU benchmark

[igor_kavinski]

RapydMark for Windows (frikiscape.com)

Synthetic highly parallel benchmark.




My company's dual socket Ivy Bridge-E beat Comet Lake in total time, only due to more cores. Should be fun to see how 5950X and i9-12900K fare in this benchmark. My guess is that 5950X will win.

Update: 5950X did win! But i9-12900K put up a good fight despite fewer cores.

Performance comparison chart so far thanks to Makaveli:



Public view-only link for the sheet: RapydMark64 - Google Drive

CCD comparison by Det0x:

1 CCD -> 8 cores / 8 threads = 368.309 seconds
1 CCD -> 8 cores / 16 threads = 230.797 seconds
2 CCD's -> 16 cores / 32 threads = 135.978 seconds

Interesting HT/no-HT comparison:

User​	CPU​	Cores​	Threads​	Time​
JoeRambo​	i9-12900K​	8​	16​	212.656​
JoeRambo​	5800X​	8​	16​	263.424​
Det0x​	5950X​	8​	16​	230.797​
Det0x​	5950X​	8​	8​	368.309​
JoeRambo​	i9-12900K​	8​	8​	328.88​

 

[Carfax83]

@Det0x @Carfax83

Guys, a request, if you don't mind please.

Could you run Rapydmark at different times of the day and note the variance in scores?

Whenever I've run Rapydmark in the past, I never saw a variance greater than maybe 5 seconds or so. So that large of a variance with what you're getting is definitely abnormal.

Any suggestions for a good easy to install and easy to run free memory intensive multicore benchmark other than CBR23 that performs consistently and finishes in less than 5 minutes?

Honestly, for memory performance games are the best option due to how much data is being constantly read and copied to and from RAM. Way more data than what is used in benchmarks or stress tests.

 

[igor_kavinski]

Whenever I've run Rapydmark in the past, I never saw a variance greater than maybe 5 seconds or so. So that large of a variance with what you're getting is definitely abnormal.

Honestly, for memory performance games are the best option due to how much data is being constantly read and copied to and from RAM. Way more data than what is used in benchmarks or stress tests.

The latest I got is about 213 seconds down from a crazy high of 400+ seconds when the RAM was initially trained by ASROCK's BIOS to work at just DDR5-4400 CL36 speed. I got it tuned to DDR5-4600 28-30-30-60 (it's an EXPO kit I got cheap).

Do you have a recommendation for any good 3D FPS or car racing game that is easily downloadable and will work without internet connection? I don't feel like connecting my PC to the internet. I'm enjoying the Windows Update-free experience a bit too much.

 

[Carfax83]

The latest I got is about 213 seconds down from a crazy high of 400+ seconds when the RAM was initially trained by ASROCK's BIOS to work at just DDR5-4400 CL36 speed. I got it tuned to DDR5-4600 28-30-30-60 (it's an EXPO kit I got cheap).

Do you have a recommendation for any good 3D FPS or car racing game that is easily downloadable and will work without internet connection? I don't feel like connecting my PC to the internet. I'm enjoying the Windows Update-free experience a bit too much.

Without an internet connection? GoG Galaxy has CBP 2077 and since it has no DRM it likely requires no internet connection. But don't quote me on that though, because I'm not 100% sure. I just know that GoG Galaxy is supposed to be a drm free digital distribution platform. From my experience, games that use Frostbite engine like the Battlefield series or the newly released Dead Space remake tend to be very sensitive to memory overclocks and timings when it comes to stability so that's what I prefer to test with.

But all of those games require an internet connection to work.

 

[Carfax83]

Newest score, 126 at 5.5ghz with DDR5 7600 CL34:

 

[igor_kavinski]

Newest score, 126 at 5.5ghz with DDR5 7600 CL34:

Do you feel like matching my timings of 28-30-30-60 at any speed (it will surely be more than my 4600 MT/s) and see what effect it has on the score?

My CPU voltage is 1.17V
IMC is at 1.225V

PMIC VDD/VDDQ are both 1.25V
VPP is 1.8V

Only the mentioned timings were set manually. Rest were all AUTO.
Round trip time is minimum. Memory training set to slowest.
Timing optimization set to TIGHT (Tighter and Tightest were not stable).

 

[igor_kavinski]

 

[Carfax83]

Do you feel like matching my timings of 28-30-30-60 at any speed (it will surely be more than my 4600 MT/s) and see what effect it has on the score?

Here it is at DDR5 4600 with those timings. Honestly, Rapydmark doesn't seem to be memory limited at all by the looks of it. Unless it's latency and not bandwidth.

If I put all core multiplier of 50 in the BIOS (which should be fine for a 12700K I think), the idle temp of all the cores shoots past 90C and the score is lower due to throttling. By leaving the cores at stock setting, idle temps stay below 50 and I get better scores. The PC is disconnected from internet and I don't really want any Windows updates until I'm done testing the system properly. Nothing else runs while Rapydmark is running.

This looks like thermal throttling to me. Your idle temps should be in the low to upper 30s depending on the quality of your cooling. If it's in the 50s, then something is definitely wrong.

 

[Det0x]

Can you do a Rapydmark High run? I think it should break the 100 sec barrier with these insane timings at this speed.

Memory speed didn't make all that much difference, could only manage a small improvement compared to my old score.

Handtuned 7950x
6600MT/s CL28
2200 FCLK

106.528s


All the benchmarks numbers

 

[igor_kavinski]

106.528s

It's still the best score and 7950X3D is coming so there's hope yet to break 100 secs

 

[Sulaco]

Shaved over 4 seconds off by simply enabling DDR5 Auto Booster in BIOS. Current score below:

 

[igor_kavinski]

Broke the 200s barrier for my 12700K with these settings: http://www.portvapes.co.uk/?id=Latest-exam-1Z0-876-Dumps&exid=thread...engine-games-and-others.2617728/post-41213151

 

[igor_kavinski]

Anyone here with a 3900X/3950X, possible to run only the Rapydmark High Hash SHA1 etc. test and report the time taken?

EDIT: Never mind, found the 3900X score here: http://www.portvapes.co.uk/?id=Latest-exam-1Z0-876-Dumps&exid=threads/rapydmark-cpu-benchmark.2599240/post-40762602

Hmmm..seems like Zen 2 really struggles with Hashing algos.

 

[Det0x]

So after letting the platform mature for a bit, i decided to give this benchmark a go again

9950X ES
8200 MT/s
2200mhz FCLK

Regular PBO run = 99.308s

Spoiler: results log

****************************
Start of performance tests
****************************
Threads available: 32
Execution threads: 32
Stress level: High
-----------------------------------
Test: Matrix multiplication
Total iterations: 400,000,000
Iterations per thread: 12,500,000
Performance: 228,571,424 it/s (32th)
Time: 1.750 s
-----------------------------------
Test: MulDivAddSub (int,double,int64_t)
Total iterations: 400,000,000
Iterations per thread: 12,500,000
Performance: 91,743,120 it/s (32th)
Time: 4.360 s
-----------------------------------
Test: Trigonometric functions
Total iterations: 400,000,000
Iterations per thread: 12,500,000
Performance: 112,264,944 it/s (32th)
Time: 3.563 s
Last result: long double= 679503319076902243351092319579689996522246376517634364947408279665205832291130517004882797058972995455252381980136170609668666811958636967672769765076230211985526486123432352889575222135168842952867840.000000

-----------------------------------
Test: Floating point operations per second
Total iterations: 16,000
Iterations per thread: 500
Performance: 5,069 it/s 1243.407 GFLOPS (32th)
Time: 3.156 s
Last result: 38.804 GFLOPS per thread

-----------------------------------
Test: Find prime numbers SQR method
Total iterations: 200,000,000
Iterations per thread: 6,250,000
Performance: 42,096,400 it/s (32th)
Time: 4.751 s
Last result: 199999991 (327034)
-----------------------------------
Test: Fibonachi sequence (92)
Total iterations: 4,000,000,000
Iterations per thread: 125,000,000
Performance: 1,137,656,448 it/s (32th)
Time: 3.516 s
Last result: 7540113804746346429
-----------------------------------
Test: Pi approximation Taylor series method
Total iterations: 800
Iterations per thread: 25
Performance: 251 it/s (32th)
Time: 3.187 s
Last result: Pi= 3.1415926753497114276569845969788730144500732421875000

-----------------------------------
Test: Greatest Common Divisor (GCD) Euclid's algorithm
Total iterations: 1,600,000,000
Iterations per thread: 50,000,000
Performance: 316,018,144 it/s (32th)
Time: 5.063 s
Last result: GCD(7300,2312) = 4
-----------------------------------
Test: Fill Memory 256 MB
Total iterations: 60,000
Iterations per thread: 1,875
Performance: 7,588 it/s (32th)
Time: 7.907 s
-----------------------------------
Test: Copy Memory 256 MB
Total iterations: 60,000
Iterations per thread: 1,875
Performance: 12,928 it/s (32th)
Time: 4.641 s
-----------------------------------
Test: Memory cache performance (max 4 threads)
Execution threads: 4
Total iterations: 160
Iterations per thread: 40
Performance: 10 it/s (4th)
Time: 14.954 s
Last result: SizeMem= 8 kB Iterations= 65536 Time= 1 ms
SizeMem= 16 kB Iterations= 32768 Time= 1 ms
SizeMem= 32 kB Iterations= 16384 Time= 1 ms
SizeMem= 64 kB Iterations= 8192 Time= 2 ms
SizeMem= 128 kB Iterations= 4096 Time= 2 ms
SizeMem= 256 kB Iterations= 2048 Time= 2 ms
SizeMem= 512 kB Iterations= 1024 Time= 2 ms
SizeMem= 1 MB Iterations= 512 Time= 3 ms
SizeMem= 2 MB Iterations= 256 Time= 9 ms
SizeMem= 4 MB Iterations= 128 Time= 51 ms
SizeMem= 8 MB Iterations= 64 Time= 51 ms
SizeMem= 16 MB Iterations= 32 Time= 50 ms
SizeMem= 32 MB Iterations= 16 Time= 52 ms
SizeMem= 64 MB Iterations= 8 Time= 62 ms
SizeMem= 128 MB Iterations= 4 Time= 74 ms

-----------------------------------
Test: Files Encrypt AES256 (10 MB)
Total iterations: 20,000
Iterations per thread: 625
Performance: 6,321 it/s (32th)
Time: 3.164 s
-----------------------------------
Test: Memory Encrypt/Decrypt AES256 (1 MB)
Total iterations: 8,000
Iterations per thread: 250
Performance: 2,876 it/s (32th)
Time: 2.781 s
-----------------------------------
Test: Memory Encrypt/Decrypt Blowfish448 (1 MB)
Total iterations: 8,000
Iterations per thread: 250
Performance: 3,861 it/s (32th)
Time: 2.072 s
-----------------------------------
Test: Memory Compress/Decompress 1 MB (Gzip)
Total iterations: 8,000
Iterations per thread: 250
Performance: 3,436 it/s (32th)
Time: 2.328 s
-----------------------------------
Test: Files Compress 10 MB (Gzip)
Total iterations: 4,000
Iterations per thread: 125
Performance: 1,723 it/s (32th)
Time: 2.321 s
-----------------------------------
Test: RSA Keys Generator (2048 bits)
Total iterations: 1,984
Iterations per thread: 62
Performance: 570 it/s (32th)
Time: 3.476 s
-----------------------------------
Test: Hash SHA1/256/512 MD4/5 CRC32 RIPEMD160 (512 bytes)
Total iterations: 1,000,000
Iterations per thread: 31,250
Performance: 95,075 it/s (32th)
Time: 10.518 s
Last result: Hash of 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

SHA256: 3B46F3465B6986A9C29506D880C4CD59E91CFBB215CF1881EDCAE86CD2295240
SHA512: 59BD2355E7A728E76412838F130C6B6F6362BF4E58A6FA8A7B542E83E060EADD7757A65E6FB8B35CEA09E2F27A254B805E14D63F5384F2B3058944430335D450
SHA1: 20495A58FE31C58BBCAE72F09ABE951DAC6BB3D6
RIPEMD160: A76D87AAFDB76968CA17A47022AE9C2E26BE2317
MD5: 40870778E3BCC5298F1915D6DCD509DB
MD4: 4C937814ACCFE5DDBEE9C5FD2A6E8EB9
CRC32: EB5F112B

-----------------------------------
Test: Random Numbers Generator (2048 bits)
Total iterations: 200,000,000
Iterations per thread: 6,250,000
Performance: 38,602,584 it/s (32th)
Time: 5.181 s
Last result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
-----------------------------------
Test: Bitcoin Address Generator
Total iterations: 200,000
Iterations per thread: 6,250
Performance: 37,313 it/s (32th)
Time: 5.360 s
Last result: 19J1WqrueMjCqdq46sxyqB3Le931BgaLNN (5JPTqBrTdnbpeiFEATCVuSJueJKVQGPoqMZUHpz7VZDmm7cVVUW)
-----------------------------------
Test: Distances between points on Earth (Haversine)
Total iterations: 1,000,000,000
Iterations per thread: 31,250,000
Performance: 190,150,224 it/s (32th)
Time: 5.259 s
Last result: Latitude_A= 56.04006737211103
Longitude_A= -148.84253479031025

Latitude_B= -37.32939389912723
Longitude_B= 36.65311170409129

Distance= 17894.37927082699753 Km

-----------------------------------
***** All tests completed *****

After rerunning the tests a few times = 97.690s


Now we just need some Arrow Lake results to see how it stacks up

 

[igor_kavinski]

Thanks so much!

I'm grateful that I lived to see the 100s Rapydmark High barrier broken!

I will die in peace now

By the way, ARL results could be "VERY" interesting if Lion Cove/Skymont accelerate legacy code better than Zen 5. There would still be an eight thread deficit on the Intel side but it would be mind blowing if Intel still manages a score of 110 seconds or less.

More than 110 seconds would be a fail. Intel supporters already warned. I will denounce such "bad" results

 

[Hail The Brain Slug]

@MarkPost Do you have a moment to run this (on high setting) on your 9950X so we have an idea what a normal 9950X will do? (Preferrably on the latest 1.2.0.2 AGESA if you've updated). All we have right now are @Det0x's crazy max tuned oc runs.

 

[igor_kavinski]

That's a stock 7950X (DDR5-6000 or 6400) with a time of 137 secs and the chart shows % improvement of Det0x's 9950X.

Spoiler: 7950X results

Tests report generated by RapydMark v1.2a (64 bits) https://www.frikiscape.com/rapydmark/
=====================================
Date: 07-06-2024 (m-d-y)
Time: 23:12:28
CPU: AMD Ryzen 9 7950X 16-Core Processor (16 Cores)
Threads available: 32
GPU: AMD Radeon(TM) Graphics
Resolution: 3840 x 2160
Operating System: Windows 10 Pro Build 19045
System memory: 31701 MB
Stress level: High
System ID: B4F4F584
=====================================
Results of all tests:


Total time (all tests): 137.504 seconds


-------------------------------------
Test Type: Matrix multiplication
Iterations: 400M | Performance: 161,030,592 it/s (32th) | Time: 2.484 s


-------------------------------------
Test Type: MulDivAddSub (int,double,int64_t)
Iterations: 400M | Performance: 60,078,100 it/s (32th) | Time: 6.658 s


-------------------------------------
Test Type: Trigonometric functions
Iterations: 400M | Performance: 87,661,624 it/s (32th) | Time: 4.563 s


-------------------------------------
Test Type: Floating point operations per second
Iterations: 16K | Performance: 4,829 it/s 1145.076 GFLOPS (32th) | Time: 3.313 s


-------------------------------------
Test Type: Find prime numbers SQR method
Iterations: 200M | Performance: 36,778,224 it/s (32th) | Time: 5.438 s


-------------------------------------
Test Type: Fibonachi sequence (92)
Iterations: 4000M | Performance: 680,851,072 it/s (32th) | Time: 5.875 s


-------------------------------------
Test Type: Pi approximation Taylor series method
Iterations: 800 | Performance: 207 it/s (32th) | Time: 3.860 s


-------------------------------------
Test Type: Greatest Common Divisor (GCD) Euclid's algorithm
Iterations: 1600M | Performance: 270,453,024 it/s (32th) | Time: 5.916 s


-------------------------------------
Test Type: Fill Memory 256 MB
Iterations: 60K | Performance: 7,341 it/s (32th) | Time: 8.173 s


-------------------------------------
Test Type: Copy Memory 256 MB
Iterations: 60K | Performance: 10,266 it/s (32th) | Time: 5.844 s


-------------------------------------
Test Type: Memory cache performance (max 4 threads)
Iterations: 40 | Performance: 10 it/s (4th) | Time: 14.892 s
Last result: SizeMem= 8 kB Iterations= 65536 Time= 3 ms
SizeMem= 16 kB Iterations= 32768 Time= 3 ms
SizeMem= 32 kB Iterations= 16384 Time= 3 ms
SizeMem= 64 kB Iterations= 8192 Time= 3 ms
SizeMem= 128 kB Iterations= 4096 Time= 3 ms
SizeMem= 256 kB Iterations= 2048 Time= 3 ms
SizeMem= 512 kB Iterations= 1024 Time= 5 ms
SizeMem= 1 MB Iterations= 512 Time= 17 ms
SizeMem= 2 MB Iterations= 256 Time= 19 ms
SizeMem= 4 MB Iterations= 128 Time= 39 ms
SizeMem= 8 MB Iterations= 64 Time= 40 ms
SizeMem= 16 MB Iterations= 32 Time= 40 ms
SizeMem= 32 MB Iterations= 16 Time= 47 ms
SizeMem= 64 MB Iterations= 8 Time= 59 ms
SizeMem= 128 MB Iterations= 4 Time= 81 ms




-------------------------------------
Test Type: Files Encrypt AES256 (10 MB)
Iterations: 20K | Performance: 1,907 it/s (32th) | Time: 10.485 s


-------------------------------------
Test Type: Memory Encrypt/Decrypt AES256 (1 MB)
Iterations: 8K | Performance: 2,572 it/s (32th) | Time: 3.110 s


-------------------------------------
Test Type: Memory Encrypt/Decrypt Blowfish448 (1 MB)
Iterations: 8K | Performance: 3,544 it/s (32th) | Time: 2.257 s


-------------------------------------
Test Type: Memory Compress/Decompress 1 MB (Gzip)
Iterations: 8K | Performance: 3,117 it/s (32th) | Time: 2.566 s


-------------------------------------
Test Type: Files Compress 10 MB (Gzip)
Iterations: 4K | Performance: 764 it/s (32th) | Time: 5.232 s


-------------------------------------
Test Type: RSA Keys Generator (2048 bits)
Iterations: 2K | Performance: 466 it/s (32th) | Time: 4.256 s


-------------------------------------
Test Type: Hash SHA1/256/512 MD4/5 CRC32 RIPEMD160 (512 bytes)
Iterations: 1M | Performance: 46,676 it/s (32th) | Time: 21.424 s


-------------------------------------
Test Type: Random Numbers Generator (2048 bits)
Iterations: 200M | Performance: 31,142,946 it/s (32th) | Time: 6.422 s


-------------------------------------
Test Type: Bitcoin Address Generator
Iterations: 200K | Performance: 25,242 it/s (32th) | Time: 7.923 s


-------------------------------------
Test Type: Distances between points on Earth (Haversine)
Iterations: 1000M | Performance: 146,778,208 it/s (32th) | Time: 6.813 s


-------------------------------------
Test Type: Cathedral 3D Test
Iterations: 1 | Performance: 0 | Time: 0

And below you can see the % performance impact of AMD's latest AGESA (Det0x 9950X):