Low 2D Performance scores

[kapqa]

i am having a similar problem also with 2D Graphics Mark Score.

with a recent Ryzen System and a GeForce 1650 on Windows 7-64Bit

with Passmark Version 9 i have around a Passmark Rating of 6,700 and a 2D Graphics Mark of 988.

however, upon updating to latest Passmark Version 10, the same system gets an abyssmaly low Passmark of 333 and a 2D Graphics Mark of 22,


so i really wonder what is going on here.
For the Nvidia part, i have installed the quite latest drivers (461,09 for Windows 7)

[Richard (PassMark)]

i am having a similar problem also with 2D Graphics Mark Score.

with a recent Ryzen System and a GeForce 1650 on Windows 7-64Bit

with Passmark Version 9 i have around a Passmark Rating of 6,700 and a 2D Graphics Mark of 988.

however, upon updating to latest Passmark Version 10, the same system gets an abyssmaly low Passmark of 333 and a 2D Graphics Mark of 22,


so i really wonder what is going on here.
For the Nvidia part, i have installed the quite latest drivers (461,09 for Windows 7)

Per the FAQ, "2D results are low for Windows 7 when using PerformanceTest V10?",

Some test scores and the 2D mark value in older versions of Windows (eg Windows 7) will be far lower in PerformanceTest V10 as it isn't possible to scale meaningfully when multiple tests can't be run.

Newer versions of Windows have added more available drawing APIs such as PDF and SVG rendering functions. The PDF and SVG test will not be able to run in Windows 7 at all due to missing Windows API functions. The DX11 parts of the Image Filters and Image Rendering test may be able to run in Windows 7 if DX11 is installed and a DX11 compatible video card is being used.

[kapqa]

Thank you for the explanation; i ran the test also on another computer with AMD APU A10-9700, and altough the difference is not so big between v9 and v10, it is still noticeable.
So i am wondering now why the Graphics score is so much lower for the Nvidia Card, which in Theory should outperform the AMD APU Radeon R7 many times.
Could it be that the card itself is somehow faulty or is it more related to the testing methodology/Nvidia - Windows 7 driver issues?

[David (PassMark)]

kapqa,

Have a look at this post. It covers all the common performance issues
Causes and Solutions for a slow PC

[RobertB]

Sorry for the late acknowledgement - I've been away for a few days.
I am definitely on the trailing edge of technology - budget, and I like to fiddle about with the old stuff at a 90% discount. And solve mysteries if I can.

[Radi_SVK]

Hello, I am in a same position as OP. Except I have a Top -end machine in tip top state from stability to maintenance.
I have a 5800X machine running with 32GB of 3600MHz ram and as for the the point, or should I say on the issue.. I have an RTX 3080. Naturally, my system is way up there in the scores, but the 2D score is extremely low, 180 I think last time I checked.. Interested in the solution.

[Richard (PassMark)]

Hello, I am in a same position as OP. Except I have a Top -end machine in tip top state from stability to maintenance.
I have a 5800X machine running with 32GB of 3600MHz ram and as for the the point, or should I say on the issue.. I have an RTX 3080. Naturally, my system is way up there in the scores, but the 2D score is extremely low, 180 I think last time I checked.. Interested in the solution.

Within PerformanceTest's Manage Baseline feature, download and search for other submitted benchmarks with similar specs as yours and compare. The 2D Mark is computed from the individual tests, so see if your individual 2D scores are reasonable with others baselines.

If you think it really is a problem, what are the individual scores for the 2D test.

[Radi_SVK]

Within PerformanceTest's Manage Baseline feature, download and search for other submitted benchmarks with similar specs as yours and compare. The 2D Mark is computed from the individual tests, so see if your individual 2D scores are reasonable with others baselines.

If you think it really is a problem, what are the individual scores for the 2D test.

Hello Richard, thanks for that tip on filtering results. It was really helpful as I was able to filter results exactly for my CPU + GPU.. and while I did run just now extensive re-run of the 2D test will all kind of settings and everything is back to normal.. I was not able to determine the previous low score.. I am sure it was only momentary, perhaps something was running in the background that eats up resources.

[HHH03]

The 2D test has been poison for me ever since the Spectre/Meltdown patches came out. Unfortunately, these patches and AMD has made my rig very “outdated”. The 2D test can really reduce an awesome test score with a low 2D result... the only way I can achieve decent 2D scores is if I set Windows for “Adjust for best performance” and disable the Spectre patch. For some reason Meltdown patch doesn’t slow things up much if at all. The 3D performance is often better when Windows is set for “Adjust for best appearance” and Spectre/Meltdown patches are enabled.

[Aelfy]

I'm having the same problem of the incredibly low 2D score on the PDF test (in the 1st percentile). This is a with a brand new Dell G5 Gaming Desktop with i5-10400F, GTX 1660 Ti video card, 16GB ram, SSD. All the other scores are great. The PDF score can't be correct. I opened the TestPDF.pdf file and it scrolls through no problem using Acrobat. I tried disabling stuff with the inspectre tool mentioned above, but it didn't fix it. I tried installing the DX10 that someone suggested, but it didn't work. My windows 10 build is completely updated with Windows Update. I am using the most recent NVIDIA game ready driver ver 461.72. Now I'm worrying there could be something wrong with my computer - almost wish I didn't run the performance test. Any ideas?

[David (PassMark)]

Aelfy, it is very unlikely it is a hardware problem. More likely it is a software / configuration issue.

When you run the 2D PDF render test in PerformanceTest does it look slow? (i.e. does the benchmark score correspond to the visual experience).

On a average machine the pages should flash past. At least 5 to 10 pages a second.

[Aelfy]

I do see the 2d PDF render - It looks like 4 short bursts of scrolling, each followed by 1-2sec of pausing on a page, with the total of the time for the whole PDF test lasting maybe 6-7 seconds.

[David (PassMark)]

There is no scrolling in the PDF test. It displays each page one after the other, but not with scrolling as you would see with the scroll bar. There should not be any pause. So if it was paused for 4 seconds of a 6 second test, that would explain the low result and also prove it isn't a calculation error (i.e. it seems to count the number of pages rendered correctly).

Not sure why it is pausing. But from the 1000+ results we have for the 1660TI video card, no one else reported results under around 25 in the PDF render test. So it isn't a common issue.

[Aelfy]

There is no scrolling in the PDF test. It displays each page one after the other, but not with scrolling as you would see with the scroll bar. There should not be any pause. So if it was paused for 4 seconds of a 6 second test, that would explain the low result and also prove it isn't a calculation error (i.e. it seems to count the number of pages rendered correctly).

Not sure why it is pausing. But from the 1000+ results we have for the 1660TI video card, no one else reported results under around 25 in the PDF render test. So it isn't a common issue.

Okay I have some useful information after I downgraded the NVIDIA driver. I completely uninstalled all NVIDIA components of their most recent game ready driver 461.72. I then installed the oldest NVIDIA driver (driver only) I could find on their website for my 1660 ti - game ready version 457.30. All of the 2D graphics tests were higher and the PDF test was MUCH higher (54, 92nd percentile). The overall test result was much higher as well at 6593 (note this is with a +100 core overclock and +1000 memory overclock).

So, to summarize, something is wonky with the latest NVIDIA driver 461.72 (game ready). Downgrading to 457.30 (game ready) worked better with Passmark Performance Test, yielding a reasonable score.

[David (PassMark)]

Aelfy,
Thanks for follow up. Video card device drivers have been like this for decades, each release improves a few things, but breaks other things. I suspect the vendors are in such as rush to do each release (to line up with hardware releases) that they never do proper testing. Still it does seem to be a rare issue, so maybe it is only a problem when used in combination with something else (particular hardware, or overclocking settings, or some 3rd party software).