MEMORY MARK - Database Operations - Very low score

[DTX2023]

Hi,
I've recently built a new workstation based on a dual AMD EPYC 7763 with 2TB of DDR4-3200MHz RAM and I'm getting 1/10 of the score (for MEMORY Database Operations) I'm getting on my older dual XEON 6258R with 768GB of DDR4-2933 MHz.
Database Operations score on the dual AMD 7763 is 966
Database Operations score on the dual XEON 6258R is 9160 +
The rest of the memory mark is fine.
Any idea what could be causing this slow performance? Any BIOS or Windows setting I should be looking for?
Thanks.

[David (PassMark)]

Everything below is for PerformanceTest V10 and V11. The memory test hasn't really changed between these releases.

We had a report earlier this week about this that was similar.
But we don't know the cause at the moment.

Here is what we do know.

For a certain CPU model (any CPU) there are a range of different performance level. The differences are explained by overclocking, different memory speeds, single or dual channel setups, BIOS timing configs, background software running on the machine, single rank RAM or dual rank, ECC RAM or not, buffered RAM or not, etc... etc....

So the distribution of results for the database test typically looks like this (around 1200 different machines with the Intel 13600K CPU for this example graph).
There are a few machines (a couple of percent of all machines tested) at both extremes. A few are really really messed up and a few are highly optimised. The vast majority sit somewhere in the middle.



However for a small number of high core count CPUs we see an excess of machines at the low end. Instead of seeing maybe 3% of machines being in a bad messed up state, we are seeing more around 10% to 15%.

This 10% number seems to be too high to be explained by just bad settings however. On the other hand these high core count CPUs are much more difficult to setup in an optimal fashion. For example they tend to have lots of RAM slots, use multiple memory busses and have NUMA issues. So it is much easier, for example, to end up in single channel mode accidentally. But bad setup should affect more than just the database test and in at some cases we don't see that. Just the database test is affected.

Here is a graph of the Intel 13900K results for the database test and the Threaded memory test.

There is good correlation (as expected) between the database test and the multi-threaded test. Both tests should be pushing the RAM and CPU's memory controller to the limit.
But there is significant de-correlation for the slow database results for this 13900K CPU that we don't see for CPUs with smaller core counts.


2nd post to follow with some suggestions.

[David (PassMark)]

Some suggestions to investigate

1) Run the test a few times to see how random the result is.
2) Make sure the machine is as clean as possible (no significant background software running)
3) Check the RAM config is optional (best slots for each socket, best timings, not mixed RAM, etc..)
4) Collect the debug logs from PerformanceTest.
5) Try running the test with a forced lower process count. e.g. 4, 8, 16, 32 processes only. That dual EPYC 7763 you have will have 256 threads. It isn't unusual at all to see performance drop significantly once the RAM bandwidth is maxed out. You can set the process count from the preferences window.

Report back what you find out.

[ChurchillBell]

Does anyone have any tips or recommendations for optimizing database operations and increasing Memory Mark? I would like to hear about your experiences and strategies to improve performance in this area.

[DTX2023]

Hi David,
Thanks for your quick reply.
Answers below.

Some suggestions to investigate

1) Run the test a few times to see how random the result is.
- Ran the test at least 10 times and got the same low results
2) Make sure the machine is as clean as possible (no significant background software running)
- Nothing running in the background. I monitor processes and resources all the time.
3) Check the RAM config is optional (best slots for each socket, best timings, not mixed RAM, etc..)
- I'm using all 16x DIMM slots on my MB. No mixed RAM. All Samsung 128GB Reg-ECC DIMMs.
4) Collect the debug logs from PerformanceTest.
- Will do that later.
5) Try running the test with a forced lower process count. e.g. 4, 8, 16, 32 processes only. That dual EPYC 7763 you have will have 256 threads. It isn't unusual at all to see performance drop significantly once the RAM bandwidth is maxed out. You can set the process count from the preferences window.
- I got up to ~1500 score on db operations with 8 process count. After that, it drops back to ~900

My main application is Altair Feko in which I perform mainly matrix operations (inversions).
I'm getting much lower performance with my Dual EPYC on long simulations versus my Dual Xeon machine (56 Core total and 768GB of RAM) thus the reason why I want to investigate these low numbers in the database operations. Based on your experience, would the database operations performed in passmark have any correlation with matrix operations (inversions)?

Thanks again.

[David (PassMark)]

128GB Reg-ECC DIMMs

Registered ECC RAM is slower than normal desktop RAM, but that in no way explains a benchmark result of just 966.

I got up to ~1500 score on db operations with 8 process count

Even 1500 isn't great. The peak performance is at a very low core count however. Would be a shame if the whole system was bottlenecked at 8 cores when you have 256 available.

More things to check

6) Set the test duration to long in the Preferences window as this give you more time to check stuff. Then run the test a few times. While the test is running check a) CPU and system temperatures and b) clock speeds on the CPUs, using CPU-Z or task manager. Make sure the CPUs aren't getting close to being thermally throttled. Also make sure the CPU load if sufficient to get the CPUs ramping up to their max clock speed. Maybe the CPUs don't see enough load, as they are waiting for the memory bus and then most of the CPU cores clock down to idle speeds.

7) Check that no other 3rd party process spikes up at the same time as the test is running (e.g. disk, GPU or CPU load). e.g. Some security solution becomes super active when PerformanceTest launches 100+ threads.

If easy, can you disable a whole CPU socket in BIOS? This helps mitigate some of the NUMA issues with multiple memory busses.

9) Confirm with CPU-Z that the memory controller has placed the RAM into 8 channel mode. Maybe BIOS (wrongly) detected something difference in the RAM sticks and you are in single channel mode.

Based on your experience, would the database operations performed in passmark have any correlation with matrix operations (inversions)?

Fast matrix operations are often done using vector instructions (SIMD / SSE). The extended instruction test in the CPU suite in PerformanceTest actually does this with matrix multiplications. But there are many ways to code this and different CPUs have different accelerated instructions. Maybe the XEON code path was better than the EPYC code path for your case (i.e. your code wasn't optimised / accelerated for the new CPU). But it could also be a memory bandwidth issue and thus maybe have a link with the poor DB benchmark result. In short, it is difficult to be sure of anything without more detailed investigation.

[Justanotherguy]

Hi there,

same issue. Works as expected when just booted up, after running for a while doing the things I do on a computer it shows differently. See below. There's no change in perfomance - no change in either games or realtime audio processing which are the 2 most demanding things I do.

13900kf
32GB Corsair DDR5-6000
Windows 11

When I perform a test after a fresh reboot:


When I perform a test after a while:

[David (PassMark)]

Yes, there is definitely something strange with the database test and the 13900K CPU (see the graphs in the 2nd post), which we don't see in most other CPUs. Around 15% of 13900K systems seems messed up some of the time.

If you could follow the same debug steps above that would be good.

We are assuming the original poster resolved their issue somehow.

[Justanotherguy]

Some suggestions to investigate

1) Run the test a few times to see how random the result is.
2) Make sure the machine is as clean as possible (no significant background software running)
3) Check the RAM config is optional (best slots for each socket, best timings, not mixed RAM, etc..)
4) Collect the debug logs from PerformanceTest.
5) Try running the test with a forced lower process count. e.g. 4, 8, 16, 32 processes only. That dual EPYC 7763 you have will have 256 threads. It isn't unusual at all to see performance drop significantly once the RAM bandwidth is maxed out. You can set the process count from the preferences window.

Report back what you find out.

1) Same result - when it is a low score it keeps being a low score until restart.
2) It is 1 month old, pretty fresh install
3) 1 x 32GB
4) See below
5) It was 32, tried changing it to 16 with no change.

With 32 process count:

0.000s - DEBUG: Starting...
0.000s - DEBUG: Start time: 2023/05/28, 12:55:47.758
0.000s - DEBUG: PerformanceTest 10.2 build 1017 64-Bit
0.000s - DEBUG OS: Windows 11 Home build 22621 (64-bit)
0.000s - DEBUG Path: C:\Program Files\PerformanceTest
0.000s - DEBUG Command line: "C:\Program Files\PerformanceTest\PerformanceTest64.exe" debugmode
0.000s - Date: 05/28/23 12:55:47
0.000s - GetPersonalFolder: C:\Users\formi\Documents
0.000s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
0.000s - Selected drive C
0.000s - HORZRES - 3440 VERTRES - 1440
0.000s - LOGPIXELSY - 96 LOGPIXELSX - 96 DPIScale - 1.000000
0.000s - TEMP DEBUG csum ok: 1
0.000s - DEBUG: Opening splash window
0.016s - GetCommonApplicationDataFolder: C:\ProgramData
0.047s - DEBUG: Retrieving Computer Name
0.047s - DEBUG: Retrieving OS Name
0.062s - Windows 11 Home build 22621 (64-bit)
0.062s - DEBUG: Retrieving Graphics Info
0.328s - DEBUG: SysInfoDll_GetGraphicsInfo OK
0.328s - Video adapters: check unsupported
0.328s - Video adapters: iNumDD 1
0.328s - Video adapter 0: bDisplay: 1 szDeviceDescription: NVIDIA GeForce RTX 4070 Ti szDriverVersion: 531.61 : Memory: 12878086144
0.328s - Monitor 0:1 3440x1440x32 144Hz 96 DPI (Primary monitor)
0.328s - DEBUG: Retrieving USB Info
0.437s - DEBUG: GetProcessList
0.437s - DEBUG: SysInfoDLL_Results::GetProcessList start
0.437s - DEBUG: SysInfoDLL_Results::GetProcessList finished
0.437s - DEBUG: Get apps
0.469s - DEBUG: finished quickgather
0.469s - DEBUG: Checking language and local
0.469s - DEBUG: Got language and local
0.469s - DEBUG: Creating main window
0.484s - DEBUG: Updating Baseline List
0.484s - Updating baseline: using OS:
0.484s - Windows 11 Home build 22621 (64-bit)
0.500s - DEBUG: Finished updating Baseline List
0.516s - Got WM_EXITSIZEMOVE message
0.531s - CD3DApplication::BuildDeviceList start
0.531s - CD3DApplication::BuildDeviceList adapter 0 of 1
0.531s - CD3DApplication::BuildDeviceList finish
0.531s - DEBUG PERF: Initialize3DEnvironment
1.484s - g_D3D9ComplexDeviceUsed found (NVIDIA GeForce RTX 4070 Ti)
1.484s - Got WM_EXITSIZEMOVE message
1.484s - WM_EXITSIZEMOVE: BackBufferWidth 690 : BackBufferHeight 434 (previous: 491 : 625)
1.500s - Got WM_EXITSIZEMOVE message
1.500s - WM_EXITSIZEMOVE: BackBufferWidth 674 : BackBufferHeight 568 (previous: 690 : 434)
1.500s - GetPersonalFolder: C:\Users\formi\Documents
1.500s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
1.500s - GetPersonalFolder: C:\Users\formi\Documents
1.500s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
1.500s - GetPersonalFolder: C:\Users\formi\Documents
1.500s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
1.500s - GetPersonalFolder: C:\Users\formi\Documents
1.500s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
1.500s - GetPersonalFolder: C:\Users\formi\Documents
1.500s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
1.500s - GetPersonalFolder: C:\Users\formi\Documents
1.500s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
1.500s - GetPersonalFolder: C:\Users\formi\Documents
1.500s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
1.500s - GetPersonalFolder: C:\Users\formi\Documents
1.500s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
1.500s - GetPersonalFolder: C:\Users\formi\Documents
1.500s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
1.500s - GetPersonalFolder: C:\Users\formi\Documents
1.500s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
1.500s - DEBUG: Creating performance stats thread
1.500s - DEBUG PERF: CPerformanceStats::Init start
1.625s - DEBUG PERF: CPerformanceStats PdhAddCounter failed -1073738823
1.625s - DEBUG: Performance stats initialised
1.625s - DEBUG: Initialising bitmap resources
1.625s - DEBUG: Closing splash window
1.625s - DEBUG: RP 1
1.625s - DEBUG: RP 2
1.625s - DEBUG: RP 3
1.625s - DEBUG: RP 4
1.625s - DEBUG: RP 5
1.625s - DEBUG: RP 6
1.703s - DEBUG: RP 7
1.703s - DEBUG: RP 9
1.703s - DEBUG: RP 11
1.719s - DEBUG: RP 12
4.328s - DEBUG: RP 14
4.328s - DEBUG: RP 15
4.328s - DEBUG: RP 18
4.328s - DEBUG: Finished Welcome Window
4.391s - DEBUG: Launch System Information Thread
4.391s - DEBUG: Check for battery power
4.391s - DEBUG: Check for power saver plan
4.391s - DEBUG: Check for script
4.391s - DEBUG: Entering message loop
4.484s - SysInfoThread start
4.484s - DEBUG: SlowGather
4.484s - DEBUG: Retrieving Disk Info
5.047s - GetCommonApplicationDataFolder: C:\ProgramData
5.047s - GetChartDataFromCache - C:\ProgramData\PassMark\PerformanceTest10\Chart Data\meRating\all.xml
5.125s - DEBUG: Retrieving SMART Info
6.469s - DEBUG: Retrieving Memory Info
6.469s - DEBUG: Retrieving CPU Info
6.891s - GetCommonApplicationDataFolder: C:\ProgramData
6.891s - GetChartDataFromCache - C:\ProgramData\PassMark\PerformanceTest10\Chart Data\ME_ALLOC_S\all.xml
10.422s - GetCommonApplicationDataFolder: C:\ProgramData
10.422s - GetChartDataFromCache - C:\ProgramData\PassMark\PerformanceTest10\Chart Data\ME_ALLOC_S\all.xml
12.031s - AMD integrated graphics check
12.031s - [DD 0] 0x2782
12.031s - DEBUG: Retrieving Memory SPD Info
13.578s - DEBUG: Retrieving WMI Info
14.109s - DEBUG: Retrieving SMBIOS Info
15.937s - DEBUG: SysInfoDLL failed to get SMBIOS Info
16.000s - DEBUG: Disk Info
16.016s - Hard Drive: KINGSTON SFYRD2000G (2TB) NVMe
16.016s - Hard Drive: CT4000P3PSSD8 (4TB) NVMe
16.016s - DEBUG: SlowGather - Get temperatures
17.422s - DEBUG: SlowGather - finished
17.422s - 2ysInfoThread GetCleanModelName CPU (13th Gen Intel(R) Core(TM) i9-13900KFd)
17.422s - GetCleanModelName GetModelListFromCache
17.422s - GetCommonApplicationDataFolder: C:\ProgramData
17.422s - GetCleanModelName find clean name
17.422s - GetCleanModelName Intel Core i9-13900KF
17.422s - GetCleanModelName sort model list
17.422s - GetCleanModelName clean up list
17.422s - GetCleanModelName finished 1 (using Intel Core i9-13900KF)
17.422s - SysInfoThread GetCleanModelName G2D (NVIDIA GeForce RTX 4070 Ti)
17.422s - SysInfoThread NVIDIA GeForce RTX 4070 Ti
17.422s - GetCleanModelName GetModelListFromCache
17.422s - GetCommonApplicationDataFolder: C:\ProgramData
17.422s - GetCleanModelName find clean name
17.422s - GetCleanModelName GeForce RTX 4070 Ti
17.422s - GetCleanModelName sort model list
17.422s - GetCleanModelName clean up list
17.422s - GetCleanModelName finished 1 (using GeForce RTX 4070 Ti)
17.422s - SysInfoThread GetCleanModelName Mem
17.422s - GetCleanModelName GetModelListFromCache
17.422s - GetCommonApplicationDataFolder: C:\ProgramData
17.437s - GetCleanModelName find clean name
17.437s - GetCleanModelName sort model list
17.437s - GetCleanModelName clean up list
17.437s - GetCleanModelName finished 0 (using )
17.437s - SysInfoThread GetCleanModelName HDD
17.437s - GetCommonApplicationDataFolder: C:\ProgramData
17.453s - SysInfoThread GetCleanModelName HDD g_cleanHDDModelList numItems: 2
17.453s - KINGSTON FURY Renegade 2TB
17.453s - CT4000P3PSSD8
17.453s - SysInfoThread GetCleanModelName HDD SMARTDeviceInfo
17.453s - KINGSTON SFYRD2000G
17.453s - SysInfoThread GetCleanModelName HDD g_szCleanHDDModel KINGSTON SFYRD2000G iDiskIndex 0
17.453s - KINGSTON FURY Renegade 2TB
17.453s - SysInfoThread GetCleanModelName HDD OK
17.453s - SysInfoThread Redraw
17.453s - SysInfoThread finished
17.562s - DEBUG: Updating Baseline List
17.562s - Updating baseline: using OS:
17.562s - Windows 11 Home build 22621 (64-bit)
17.578s - DEBUG: Finished updating Baseline List
17.594s - GetCommonApplicationDataFolder: C:\ProgramData
17.594s - GetChartDataFromCache - C:\ProgramData\PassMark\PerformanceTest10\Chart Data\ME_ALLOC_S\all.xml
19.281s - GetCommonApplicationDataFolder: C:\ProgramData
19.281s - GetChartDataFromCache - C:\ProgramData\PassMark\PerformanceTest10\Chart Data\ME_ALLOC_S\all.xml
20.734s - DEBUG: Running Test - Database Operations
+0.000s - DEBUG: Starting logging...PT-DatabaseTest64.exe
+0.000s - DEBUG Path: C:\Program Files\PerformanceTest
+0.000s - DEBUG Command line: "C:\Program Files\PerformanceTest\PT-DatabaseTest64.exe"
+0.000s - Date: 05/28/23 12:56:08
+0.000s - DEBUG test_mem_sqlite_threaded : start 24 threads
+0.000s - DEBUG test_mem_sqlite_threaded : groupCount 1 maxThreads 32 step 1 offset 0
+0.000s - DEBUG test_mem_sqlite_threaded : thread 0 - group 0 mask 1
+0.000s - DEBUG test_mem_sqlite_threaded : thread 1 - group 0 mask 2
+0.000s - DEBUG test_mem_sqlite_threaded : thread 2 - group 0 mask 4
+0.000s - DEBUG test_mem_sqlite_threaded : thread 3 - group 0 mask 8
+0.000s - DEBUG test_mem_sqlite_threaded : thread 4 - group 0 mask 16
+0.000s - DEBUG test_mem_sqlite_threaded : thread 5 - group 0 mask 32
+0.000s - DEBUG test_mem_sqlite_threaded : thread 6 - group 0 mask 64
+0.000s - DEBUG test_mem_sqlite_threaded : thread 7 - group 0 mask 128
+0.000s - DEBUG test_mem_sqlite_threaded : thread 8 - group 0 mask 256
+0.000s - DEBUG test_mem_sqlite_threaded : thread 9 - group 0 mask 512
+0.000s - DEBUG test_mem_sqlite_threaded : thread 10 - group 0 mask 1024
+0.000s - DEBUG test_mem_sqlite_threaded : thread 11 - group 0 mask 2048
+0.000s - DEBUG test_mem_sqlite_threaded : thread 12 - group 0 mask 4096
+0.000s - DEBUG test_mem_sqlite_threaded : thread 13 - group 0 mask 8192
+0.000s - DEBUG test_mem_sqlite_threaded : thread 14 - group 0 mask 16384
+0.000s - DEBUG test_mem_sqlite_threaded : thread 15 - group 0 mask 32768
+0.000s - DEBUG test_mem_sqlite_threaded : thread 16 - group 0 mask 65536
+0.000s - DEBUG test_mem_sqlite_threaded : thread 17 - group 0 mask 131072
+0.000s - DEBUG test_mem_sqlite_threaded : thread 18 - group 0 mask 262144
+0.000s - DEBUG test_mem_sqlite_threaded : thread 19 - group 0 mask 524288
+0.000s - DEBUG test_mem_sqlite_threaded : thread 20 - group 0 mask 1048576
+0.000s - DEBUG test_mem_sqlite_threaded : thread 21 - group 0 mask 2097152
+0.000s - DEBUG test_mem_sqlite_threaded : thread 22 - group 0 mask 4194304
+0.000s - DEBUG test_mem_sqlite_threaded : thread 23 - group 0 mask 8388608
+0.000s - DEBUG test_mem_sqlite_threaded : wait for threads to be ready
+0.110s - DEBUG test_mem_sqlite : entry thread num 4
+0.110s - DEBUG test_mem_sqlite : entry thread num 9
+0.219s - DEBUG test_mem_sqlite : entry thread num 3
+0.235s - DEBUG test_mem_sqlite : entry thread num 8
+0.235s - DEBUG test_mem_sqlite : entry thread num 10
+0.328s - DEBUG test_mem_sqlite : entry thread num 0
+0.344s - DEBUG test_mem_sqlite : entry thread num 6
+0.453s - DEBUG test_mem_sqlite : entry thread num 7
+0.563s - DEBUG test_mem_sqlite : entry thread num 11
+0.672s - DEBUG test_mem_sqlite : entry thread num 12
+0.782s - DEBUG test_mem_sqlite : entry thread num 13
+0.891s - DEBUG test_mem_sqlite : entry thread num 14
+1.000s - DEBUG test_mem_sqlite : entry thread num 21
+1.125s - DEBUG test_mem_sqlite : entry thread num 20
+1.235s - DEBUG test_mem_sqlite : entry thread num 22
+1.344s - DEBUG test_mem_sqlite : entry thread num 23
+1.453s - DEBUG test_mem_sqlite : entry thread num 17
+1.563s - DEBUG test_mem_sqlite : entry thread num 18
+1.750s - DEBUG test_mem_sqlite_threaded : send start event
+1.750s - DEBUG test_mem_sqlite_threaded : wait for threads to complete
+16.422s - DEBUG test_mem_sqlite : results thread num 14, totalOps 218451, result 15.251762, timer diff 14323
+16.422s - DEBUG test_mem_sqlite : finished thread num 14
+16.438s - DEBUG test_mem_sqlite : results thread num 8, totalOps 218451, result 15.229434, timer diff 14344
+16.438s - DEBUG test_mem_sqlite : finished thread num 8
+16.453s - DEBUG test_mem_sqlite : results thread num 11, totalOps 218451, result 15.248569, timer diff 14326
+16.453s - DEBUG test_mem_sqlite : finished thread num 11
+16.469s - DEBUG test_mem_sqlite : results thread num 13, totalOps 218451, result 15.242186, timer diff 14332
+16.469s - DEBUG test_mem_sqlite : finished thread num 13
+16.469s - DEBUG test_mem_sqlite : results thread num 10, totalOps 218451, result 15.238996, timer diff 14335
+16.469s - DEBUG test_mem_sqlite : finished thread num 10
+16.469s - DEBUG test_mem_sqlite : results thread num 7, totalOps 218451, result 15.273088, timer diff 14303
+16.469s - DEBUG test_mem_sqlite : finished thread num 7
+16.485s - DEBUG test_mem_sqlite : results thread num 5, totalOps 218451, result 15.277362, timer diff 14299
+16.485s - DEBUG test_mem_sqlite : finished thread num 5
+16.485s - DEBUG test_mem_sqlite : results thread num 2, totalOps 218451, result 15.251762, timer diff 14323
+16.485s - DEBUG test_mem_sqlite : finished thread num 2
+16.485s - DEBUG test_mem_sqlite : results thread num 15, totalOps 218451, result 15.177587, timer diff 14393
+16.485s - DEBUG test_mem_sqlite : finished thread num 15
+16.485s - DEBUG test_mem_sqlite : results thread num 6, totalOps 218451, result 15.234745, timer diff 14339
+16.485s - DEBUG test_mem_sqlite : finished thread num 6
+16.500s - DEBUG test_mem_sqlite : results thread num 4, totalOps 218451, result 15.232620, timer diff 14341
+16.500s - DEBUG test_mem_sqlite : finished thread num 4
+16.500s - DEBUG test_mem_sqlite : results thread num 9, totalOps 218451, result 15.122950, timer diff 14445
+16.500s - DEBUG test_mem_sqlite : finished thread num 9
+16.500s - DEBUG test_mem_sqlite : results thread num 3, totalOps 218451, result 15.162838, timer diff 14407
+16.500s - DEBUG test_mem_sqlite : finished thread num 3
+16.500s - DEBUG test_mem_sqlite : results thread num 1, totalOps 218451, result 15.227311, timer diff 14346
+16.500s - DEBUG test_mem_sqlite : finished thread num 1
+16.547s - DEBUG test_mem_sqlite : results thread num 22, totalOps 218451, result 14.777177, timer diff 14783
+16.547s - DEBUG test_mem_sqlite : finished thread num 22
+16.594s - DEBUG test_mem_sqlite : results thread num 12, totalOps 218451, result 15.230495, timer diff 14343
+16.594s - DEBUG test_mem_sqlite : finished thread num 12
+16.610s - DEBUG test_mem_sqlite : results thread num 0, totalOps 218451, result 15.176532, timer diff 14394
+16.610s - DEBUG test_mem_sqlite : finished thread num 0
+16.657s - DEBUG test_mem_sqlite : results thread num 20, totalOps 218451, result 14.755218, timer diff 14805
+16.657s - DEBUG test_mem_sqlite : finished thread num 20
+16.672s - DEBUG test_mem_sqlite : results thread num 23, totalOps 218451, result 14.748244, timer diff 14812
+16.672s - DEBUG test_mem_sqlite : finished thread num 23
+16.750s - DEBUG test_mem_sqlite : results thread num 21, totalOps 218451, result 14.755218, timer diff 14805
+16.750s - DEBUG test_mem_sqlite : finished thread num 21
+17.235s - DEBUG test_mem_sqlite : results thread num 18, totalOps 218451, result 14.616996, timer diff 14945
+17.235s - DEBUG test_mem_sqlite : finished thread num 18
+17.453s - DEBUG test_mem_sqlite : results thread num 16, totalOps 218451, result 14.579924, timer diff 14983
+17.453s - DEBUG test_mem_sqlite : finished thread num 16
+17.453s - DEBUG test_mem_sqlite : results thread num 19, totalOps 218451, result 14.601363, timer diff 14961
+17.453s - DEBUG test_mem_sqlite : finished thread num 19
+17.469s - DEBUG test_mem_sqlite : results thread num 17, totalOps 218451, result 14.607222, timer diff 14955
+17.469s - DEBUG test_mem_sqlite : finished thread num 17
+17.469s - DEBUG test_mem_sqlite_threaded : complete, result 361.019592
38.203s - DEBUG PERF: calc_mark_vals CPU test 1 result is 0
38.203s - DEBUG PERF: calc mark num3DTestRun = 0
38.203s - DEBUG PERF: calc mark everything
38.203s - DEBUG PERF: numtest < 5
38.203s - DEBUG PERF: num3DTestRun 0
38.281s - TEMP DEBUG csum 2 ok: 1
38.281s - TEMP DEBUG csum 3 ok: 1
38.297s - GetCommonApplicationDataFolder: C:\ProgramData
38.297s - GetChartDataFromCache - C:\ProgramData\PassMark\PerformanceTest10\Chart Data\ME_ALLOC_S\all.xml
38.297s - GetCommonApplicationDataFolder: C:\ProgramData
38.297s - GetChartDataFromCache - C:\ProgramData\PassMark\PerformanceTest10\Chart Data\ME_ALLOC_S\all.xml
38.312s - GetCommonApplicationDataFolder: C:\ProgramData
38.312s - GetChartDataFromCache - C:\ProgramData\PassMark\PerformanceTest10\Chart Data\ME_ALLOC_S\all.xml
38.328s - GetCommonApplicationDataFolder: C:\ProgramData
38.328s - GetChartDataFromCache - C:\ProgramData\PassMark\PerformanceTest10\Chart Data\ME_ALLOC_S\all.xml
42.922s - GetApplicationDataFolder: C:\Users\formi\AppData\Local
42.922s - Saved config - Selected drive C
42.922s - DEBUG PERF: End perf startup
42.969s - DEBUG: Leaving message loop

[Justanotherguy]

With 16 process count:

0.000s - DEBUG: Starting...
0.000s - DEBUG: Start time: 2023/05/28, 12:59:15.651
0.000s - DEBUG: PerformanceTest 10.2 build 1017 64-Bit
0.000s - DEBUG OS: Windows 11 Home build 22621 (64-bit)
0.000s - DEBUG Path: C:\Program Files\PerformanceTest
0.000s - DEBUG Command line: "C:\Program Files\PerformanceTest\PerformanceTest64.exe" debugmode
0.000s - Date: 05/28/23 12:59:15
0.000s - GetPersonalFolder: C:\Users\formi\Documents
0.000s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
0.000s - Selected drive C
0.000s - HORZRES - 3440 VERTRES - 1440
0.000s - LOGPIXELSY - 96 LOGPIXELSX - 96 DPIScale - 1.000000
0.000s - TEMP DEBUG csum ok: 1
0.000s - DEBUG: Opening splash window
0.015s - GetCommonApplicationDataFolder: C:\ProgramData
0.046s - DEBUG: Retrieving Computer Name
0.046s - DEBUG: Retrieving OS Name
0.078s - Windows 11 Home build 22621 (64-bit)
0.078s - DEBUG: Retrieving Graphics Info
0.343s - DEBUG: SysInfoDll_GetGraphicsInfo OK
0.343s - Video adapters: check unsupported
0.343s - Video adapters: iNumDD 1
0.343s - Video adapter 0: bDisplay: 1 szDeviceDescription: NVIDIA GeForce RTX 4070 Ti szDriverVersion: 531.61 : Memory: 12878086144
0.343s - Monitor 0:1 3440x1440x32 144Hz 96 DPI (Primary monitor)
0.343s - DEBUG: Retrieving USB Info
0.453s - DEBUG: GetProcessList
0.453s - DEBUG: SysInfoDLL_Results::GetProcessList start
0.453s - DEBUG: SysInfoDLL_Results::GetProcessList finished
0.453s - DEBUG: Get apps
0.484s - DEBUG: finished quickgather
0.484s - DEBUG: Checking language and local
0.484s - DEBUG: Got language and local
0.484s - DEBUG: Creating main window
0.515s - DEBUG: Updating Baseline List
0.515s - Updating baseline: using OS:
0.515s - Windows 11 Home build 22621 (64-bit)
0.531s - DEBUG: Finished updating Baseline List
0.531s - Got WM_EXITSIZEMOVE message
0.546s - CD3DApplication::BuildDeviceList start
0.546s - CD3DApplication::BuildDeviceList adapter 0 of 1
0.562s - CD3DApplication::BuildDeviceList finish
0.562s - DEBUG PERF: Initialize3DEnvironment
1.562s - g_D3D9ComplexDeviceUsed found (NVIDIA GeForce RTX 4070 Ti)
1.562s - Got WM_EXITSIZEMOVE message
1.562s - WM_EXITSIZEMOVE: BackBufferWidth 690 : BackBufferHeight 434 (previous: 491 : 625)
1.578s - Got WM_EXITSIZEMOVE message
1.578s - WM_EXITSIZEMOVE: BackBufferWidth 674 : BackBufferHeight 568 (previous: 690 : 434)
1.578s - GetPersonalFolder: C:\Users\formi\Documents
1.578s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
1.578s - GetPersonalFolder: C:\Users\formi\Documents
1.578s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
1.578s - GetPersonalFolder: C:\Users\formi\Documents
1.578s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
1.578s - GetPersonalFolder: C:\Users\formi\Documents
1.578s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
1.578s - GetPersonalFolder: C:\Users\formi\Documents
1.578s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
1.578s - GetPersonalFolder: C:\Users\formi\Documents
1.578s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
1.578s - GetPersonalFolder: C:\Users\formi\Documents
1.578s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
1.578s - GetPersonalFolder: C:\Users\formi\Documents
1.578s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
1.578s - GetPersonalFolder: C:\Users\formi\Documents
1.578s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
1.578s - GetPersonalFolder: C:\Users\formi\Documents
1.578s - DEBUG GetPersonalFolder C:\Users\formi\Documents\PassMark\PerformanceTest1 0
1.578s - DEBUG: Creating performance stats thread
1.578s - DEBUG PERF: CPerformanceStats::Init start
1.703s - DEBUG PERF: CPerformanceStats PdhAddCounter failed -1073738823
1.703s - DEBUG: Performance stats initialised
1.703s - DEBUG: Initialising bitmap resources
1.703s - DEBUG: Closing splash window
1.703s - DEBUG: RP 1
1.703s - DEBUG: RP 2
1.703s - DEBUG: RP 3
1.703s - DEBUG: RP 4
1.703s - DEBUG: RP 5
1.703s - DEBUG: RP 6
1.796s - DEBUG: RP 7
1.796s - DEBUG: RP 9
1.796s - DEBUG: RP 11
1.812s - DEBUG: RP 12
4.687s - DEBUG: RP 14
4.687s - DEBUG: RP 15
4.687s - DEBUG: RP 18
4.687s - DEBUG: Finished Welcome Window
4.750s - DEBUG: Launch System Information Thread
4.750s - DEBUG: Check for battery power
4.750s - DEBUG: Check for power saver plan
4.750s - DEBUG: Check for script
4.750s - DEBUG: Entering message loop
4.843s - SysInfoThread start
4.843s - DEBUG: SlowGather
4.843s - DEBUG: Retrieving Disk Info
5.609s - DEBUG: Retrieving SMART Info
7.109s - DEBUG: Retrieving Memory Info
7.109s - DEBUG: Retrieving CPU Info
12.640s - AMD integrated graphics check
12.640s - [DD 0] 0x2782
12.640s - DEBUG: Retrieving Memory SPD Info
14.359s - DEBUG: Retrieving WMI Info
14.937s - DEBUG: Retrieving SMBIOS Info
17.015s - DEBUG: SysInfoDLL failed to get SMBIOS Info
17.109s - DEBUG: Disk Info
17.109s - Hard Drive: KINGSTON SFYRD2000G (2TB) NVMe
17.109s - Hard Drive: CT4000P3PSSD8 (4TB) NVMe
17.109s - DEBUG: SlowGather - Get temperatures
18.656s - DEBUG: SlowGather - finished
18.656s - 2ysInfoThread GetCleanModelName CPU (13th Gen Intel(R) Core(TM) i9-13900KFd)
18.656s - GetCleanModelName GetModelListFromCache
18.656s - GetCommonApplicationDataFolder: C:\ProgramData
18.671s - GetCleanModelName find clean name
18.671s - GetCleanModelName Intel Core i9-13900KF
18.671s - GetCleanModelName sort model list
18.671s - GetCleanModelName clean up list
18.671s - GetCleanModelName finished 1 (using Intel Core i9-13900KF)
18.671s - SysInfoThread GetCleanModelName G2D (NVIDIA GeForce RTX 4070 Ti)
18.671s - SysInfoThread NVIDIA GeForce RTX 4070 Ti
18.671s - GetCleanModelName GetModelListFromCache
18.671s - GetCommonApplicationDataFolder: C:\ProgramData
18.687s - GetCleanModelName find clean name
18.687s - GetCleanModelName GeForce RTX 4070 Ti
18.687s - GetCleanModelName sort model list
18.687s - GetCleanModelName clean up list
18.687s - GetCleanModelName finished 1 (using GeForce RTX 4070 Ti)
18.687s - SysInfoThread GetCleanModelName Mem
18.687s - GetCleanModelName GetModelListFromCache
18.687s - GetCommonApplicationDataFolder: C:\ProgramData
18.703s - GetCleanModelName find clean name
18.703s - GetCleanModelName sort model list
18.703s - GetCleanModelName clean up list
18.703s - GetCleanModelName finished 0 (using )
18.703s - SysInfoThread GetCleanModelName HDD
18.703s - GetCommonApplicationDataFolder: C:\ProgramData
18.718s - SysInfoThread GetCleanModelName HDD g_cleanHDDModelList numItems: 2
18.718s - KINGSTON FURY Renegade 2TB
18.718s - CT4000P3PSSD8
18.718s - SysInfoThread GetCleanModelName HDD SMARTDeviceInfo
18.718s - KINGSTON SFYRD2000G
18.718s - SysInfoThread GetCleanModelName HDD g_szCleanHDDModel KINGSTON SFYRD2000G iDiskIndex 0
18.718s - KINGSTON FURY Renegade 2TB
18.718s - SysInfoThread GetCleanModelName HDD OK
18.718s - SysInfoThread Redraw
18.718s - SysInfoThread finished
18.828s - DEBUG: Updating Baseline List
18.828s - Updating baseline: using OS:
18.828s - Windows 11 Home build 22621 (64-bit)
18.843s - DEBUG: Finished updating Baseline List
20.531s - Updating baseline: using OS:
20.531s - Windows 11 Home build 22621 (64-bit)
20.546s - DEBUG: GetTemperature
20.781s - GetApplicationDataFolder: C:\Users\formi\AppData\Local
25.546s - Updating baseline: using OS:
25.546s - Windows 11 Home build 22621 (64-bit)
25.562s - DEBUG: GetTemperature
30.546s - Updating baseline: using OS:
30.546s - Windows 11 Home build 22621 (64-bit)
30.562s - DEBUG: GetTemperature
35.546s - Updating baseline: using OS:
35.546s - Windows 11 Home build 22621 (64-bit)
35.562s - DEBUG: GetTemperature
40.546s - Updating baseline: using OS:
40.546s - Windows 11 Home build 22621 (64-bit)
40.562s - DEBUG: GetTemperature
45.546s - Updating baseline: using OS:
45.546s - Windows 11 Home build 22621 (64-bit)
45.546s - DEBUG: GetTemperature
50.546s - Updating baseline: using OS:
50.546s - Windows 11 Home build 22621 (64-bit)
50.562s - DEBUG: GetTemperature
52.906s - GetApplicationDataFolder: C:\Users\formi\AppData\Local
54.718s - GetCommonApplicationDataFolder: C:\ProgramData
54.718s - GetChartDataFromCache - C:\ProgramData\PassMark\PerformanceTest10\Chart Data\meRating\all.xml
56.484s - DEBUG: Running Test - Database Operations
+0.000s - DEBUG: Starting logging...PT-DatabaseTest64.exe
+0.000s - DEBUG Path: C:\Program Files\PerformanceTest
+0.000s - DEBUG Command line: "C:\Program Files\PerformanceTest\PT-DatabaseTest64.exe"
+0.000s - Date: 05/28/23 13:00:12
+0.000s - DEBUG test_mem_sqlite_threaded : start 16 threads
+0.000s - DEBUG test_mem_sqlite_threaded : groupCount 1 maxThreads 32 step 2 offset 1
+0.000s - DEBUG test_mem_sqlite_threaded : thread 0 - group 0 mask 2
+0.000s - DEBUG test_mem_sqlite_threaded : thread 1 - group 0 mask 8
+0.000s - DEBUG test_mem_sqlite_threaded : thread 2 - group 0 mask 32
+0.000s - DEBUG test_mem_sqlite_threaded : thread 3 - group 0 mask 128
+0.000s - DEBUG test_mem_sqlite_threaded : thread 4 - group 0 mask 512
+0.000s - DEBUG test_mem_sqlite_threaded : thread 5 - group 0 mask 2048
+0.000s - DEBUG test_mem_sqlite_threaded : thread 6 - group 0 mask 8192
+0.000s - DEBUG test_mem_sqlite_threaded : thread 7 - group 0 mask 32768
+0.000s - DEBUG test_mem_sqlite_threaded : thread 8 - group 0 mask 131072
+0.000s - DEBUG test_mem_sqlite_threaded : thread 9 - group 0 mask 524288
+0.000s - DEBUG test_mem_sqlite_threaded : thread 10 - group 0 mask 2097152
+0.016s - DEBUG test_mem_sqlite_threaded : thread 11 - group 0 mask 8388608
+0.016s - DEBUG test_mem_sqlite_threaded : thread 12 - group 0 mask 33554432
+0.016s - DEBUG test_mem_sqlite_threaded : thread 13 - group 0 mask 134217728
+0.016s - DEBUG test_mem_sqlite_threaded : thread 14 - group 0 mask 536870912
+0.016s - DEBUG test_mem_sqlite_threaded : thread 15 - group 0 mask 2147483648
+0.016s - DEBUG test_mem_sqlite_threaded : wait for threads to be ready
+0.109s - DEBUG test_mem_sqlite : entry thread num 3
+0.219s - DEBUG test_mem_sqlite : entry thread num 4
+0.328s - DEBUG test_mem_sqlite : entry thread num 5
+0.453s - DEBUG test_mem_sqlite : entry thread num 1
+0.562s - DEBUG test_mem_sqlite : entry thread num 2
+0.672s - DEBUG test_mem_sqlite : entry thread num 10
+0.781s - DEBUG test_mem_sqlite : entry thread num 11
+0.891s - DEBUG test_mem_sqlite : entry thread num 8
+1.000s - DEBUG test_mem_sqlite : entry thread num 12
+1.187s - DEBUG test_mem_sqlite_threaded : send start event
+1.187s - DEBUG test_mem_sqlite_threaded : wait for threads to complete
+9.609s - DEBUG test_mem_sqlite : results thread num 5, totalOps 218451, result 26.650118, timer diff 8197
+9.609s - DEBUG test_mem_sqlite : finished thread num 5
+9.609s - DEBUG test_mem_sqlite : results thread num 2, totalOps 218451, result 26.705500, timer diff 8180
+9.609s - DEBUG test_mem_sqlite : finished thread num 2
+9.609s - DEBUG test_mem_sqlite : results thread num 7, totalOps 218451, result 26.624132, timer diff 8205
+9.609s - DEBUG test_mem_sqlite : finished thread num 7
+9.609s - DEBUG test_mem_sqlite : results thread num 4, totalOps 218451, result 26.620886, timer diff 8206
+9.609s - DEBUG test_mem_sqlite : finished thread num 4
+9.625s - DEBUG test_mem_sqlite : results thread num 3, totalOps 218451, result 26.650118, timer diff 8197
+9.625s - DEBUG test_mem_sqlite : finished thread num 3
+9.625s - DEBUG test_mem_sqlite : results thread num 0, totalOps 218451, result 26.640368, timer diff 8200
+9.625s - DEBUG test_mem_sqlite : finished thread num 0
+9.734s - DEBUG test_mem_sqlite : results thread num 6, totalOps 218451, result 26.656622, timer diff 8195
+9.734s - DEBUG test_mem_sqlite : finished thread num 6
+9.781s - DEBUG test_mem_sqlite : results thread num 9, totalOps 218451, result 25.688030, timer diff 8504
+9.781s - DEBUG test_mem_sqlite : finished thread num 9
+9.812s - DEBUG test_mem_sqlite : results thread num 8, totalOps 218451, result 25.672935, timer diff 8509
+9.812s - DEBUG test_mem_sqlite : finished thread num 8
+9.828s - DEBUG test_mem_sqlite : results thread num 12, totalOps 218451, result 25.651833, timer diff 8516
+9.828s - DEBUG test_mem_sqlite : finished thread num 12
+9.844s - DEBUG test_mem_sqlite : results thread num 1, totalOps 218451, result 26.624132, timer diff 8205
+9.844s - DEBUG test_mem_sqlite : finished thread num 1
+9.875s - DEBUG test_mem_sqlite : results thread num 13, totalOps 218451, result 25.675951, timer diff 8508
+9.875s - DEBUG test_mem_sqlite : finished thread num 13
+9.953s - DEBUG test_mem_sqlite : results thread num 15, totalOps 218451, result 25.651833, timer diff 8516
+9.953s - DEBUG test_mem_sqlite : finished thread num 15
+9.984s - DEBUG test_mem_sqlite : results thread num 14, totalOps 218451, result 25.648821, timer diff 8517
+9.984s - DEBUG test_mem_sqlite : finished thread num 14
+10.203s - DEBUG test_mem_sqlite : results thread num 11, totalOps 218451, result 25.357052, timer diff 8615
+10.203s - DEBUG test_mem_sqlite : finished thread num 11
+10.219s - DEBUG test_mem_sqlite : results thread num 10, totalOps 218451, result 25.348225, timer diff 8618
+10.219s - DEBUG test_mem_sqlite : finished thread num 10
+10.219s - DEBUG test_mem_sqlite_threaded : complete, result 417.866547
66.703s - DEBUG PERF: calc_mark_vals CPU test 1 result is 0
66.703s - DEBUG PERF: calc mark num3DTestRun = 0
66.703s - DEBUG PERF: calc mark everything
66.703s - DEBUG PERF: numtest < 5
66.703s - DEBUG PERF: num3DTestRun 0
66.781s - TEMP DEBUG csum 2 ok: 1
66.781s - TEMP DEBUG csum 3 ok: 1
66.796s - GetCommonApplicationDataFolder: C:\ProgramData
66.796s - GetChartDataFromCache - C:\ProgramData\PassMark\PerformanceTest10\Chart Data\meRating\all.xml
66.812s - GetCommonApplicationDataFolder: C:\ProgramData
66.812s - GetChartDataFromCache - C:\ProgramData\PassMark\PerformanceTest10\Chart Data\ME_ALLOC_S\all.xml
66.812s - GetCommonApplicationDataFolder: C:\ProgramData
66.812s - GetChartDataFromCache - C:\ProgramData\PassMark\PerformanceTest10\Chart Data\ME_ALLOC_S\all.xml
66.828s - GetCommonApplicationDataFolder: C:\ProgramData
66.828s - GetChartDataFromCache - C:\ProgramData\PassMark\PerformanceTest10\Chart Data\ME_ALLOC_S\all.xml
72.718s - GetApplicationDataFolder: C:\Users\formi\AppData\Local
72.718s - Saved config - Selected drive C
72.718s - DEBUG PERF: End perf startup
72.765s - DEBUG: Leaving message loop

​

[Justanotherguy]

Maybe it is worth mentioning that Passmark doesn't recognize my ram (Corsair Vengeance XMP DDR5, C40) :

[David (PassMark)]

3) 1 x 32GB (32GB Corsair DDR5-6000)

This isn't a good memory configuration. Especially not good for a fast CPU.
You really need to be using dual channel (so two identical sticks).
Also that CPU doesn't run at 6000MT. Max is 5600MT. So 6000 is overkill.

Or are you overlocking to 6000?
Are you overlocking the CPU or RAM?

But this doesn't fully explain the problem. If it is bad, it should stay bad. Not jump between a slow and fast state

We'll have a closer look at this logs.

[Justanotherguy]

This isn't a good memory configuration. Especially not good for a fast CPU.
You really need to be using dual channel (so two identical sticks).
Also that CPU doesn't run at 6000MT. Max is 5600MT. So 6000 is overkill.

Or are you overlocking to 6000?
Are you overlocking the CPU or RAM?

But this doesn't fully explain the problem. If it is bad, it should stay bad. Not jump between a slow and fast state

We'll have a closer look at this logs.

Another stick is on the way.

6000 is the XMP profile.

No (other than XMP profile) overclocking.

[keith]

Maybe it is worth mentioning that Passmark doesn't recognize my ram (Corsair Vengeance XMP DDR5, C40) :

Can you also attach a copy of the "SysInfoLog.txt" file as well (should be in the same directory as "PerfTestLog.txt")

[Simon (PassMark)]

@Justanotherguy

Can you open Task Manager > Performance Tab > CPU (and if not set, right click graph > Change Graph to > Logical Processors), run the Database Operations test and check that usage is spread across a lot of threads and not running on just a few/a single thread? (a screenshot of the CPU usage graph right after running the database test would help)

Then can you try running the two standalone versions of the database test here: https://passmark.com/downloads/temp/...Standalone.zip and send us the generated log files

One of the standalone versions has "processor group" (PG) and "affinity" code removed. PG and affinity lock software threads to cores. We are wondering if this causes a problem on some machines.