Copy link to clipboard
Copied
I have just done the standard Puget test but I have nothing to compare it to so I don't know if it is good or bad. Has anyone else with a similar system done the test.
Windows 10
Intel i9-9900K
64 GB RAM
RTX 2070 Super
1Tb Samsung SSD 860 boot drive
1Tb Samsung M2 projects and Export
500 Gb Samsung M2 Cache
Puget Results:
Standard Overall 623
Standard Live Playback 53.8
Standard Export 70.8
GPU Score 52
Copy link to clipboard
Copied
That system should have significantly beaten my 617 average score of my AMD Ryzen R7 3800X system, if you had Intel's QuickSync enabled. That would require force-enabling the integrated Intel UHD Graphics in your computer system's BIOS. Although granted, my system has only 32 GB of RAM, a lesser RTX 2060 SUPER GPU and only a 500 GB Samsung 970 EVO Plus m.2 NVMe SSD for my projects and export and a Samsung 850 PRO SATA SSD for my cache drive (my boot drive is only a 250 GB Samsung 850 EVO SATA SSD). Without QuickSync, that i9-9900K system of yours is a wash (or effectively a draw) against my more humble 3800X rig.
As it stands now, the lack of QuickSync dragged down your live playback score (and to a lesser extent your overall export score) while your system's GPU score from that RTX 2070 SUPER is lower than what I have achieved with my lesser RTX 2060 SUPER GPU. Those results indicate that your system needs to be tuned properly (whereas mine is close to properly tuned, without any special tweaks to the configuration).
Copy link to clipboard
Copied
Here are my Puget Results:
Standard Overall 603
Standard Live Playback 64.6
Standard Export 56
GPU Score 58
This is the first run after a Windows update. The export score may be hindered by the relatively low amount of system RAM.
Copy link to clipboard
Copied
Could you point me to a system tuning resource . I'm sure I have too many processes running but I'm not sure which I can safely delete. I have hardware encoding available as an option for H264 exports so I think that is working.
Copy link to clipboard
Copied
Puget Results (Second Run):
Standard Overall 614
Standard Live Playback 66.2
Standard Export 56.6
GPU Score 58.3
Will report results of the third run, and then average the results, in a future post.
Copy link to clipboard
Copied
Third Run:
Standard Overall 609
Standard Live Export 65.3
Standard Export 56.5
GPU Score 58.3
Averaging these results together, I now get this:
Standard Overall 609
Standard Live Playback 65.4
Standard Export 56.4
GPU Score 58.2
Still, I believe that there is a bottleneck somewhere in your system. How could an RTX 2060 SUPER outperform an ostensibly more powerful RTX 2070 SUPER?
Copy link to clipboard
Copied
Upon my further investigation into the discussion starter's GPU score, that score is more like a non-SUPER RTX 2060. That RTX 2070 SUPER should have scored in the low-60s in the GPU score.
And my particular RTX 2060 SUPER, on the other hand, scored higher than it normally would have due to most third-party manufacturing partners actually apply an overclock to the GPU clocks at the factory. My RTX 2060 SUPER is certainly one of those cards whose GPU itself has been overclocked at the factory.
Copy link to clipboard
Copied
what's a puget test ??
Copy link to clipboard
Copied
It's a series of tests produced by a PC supplier for bench marking computers for a number of different softwares. This is a link to the Premiere test:
https://www.pugetsystems.com/labs/articles/PugetBench-for-Premiere-Pro-1519/
Copy link to clipboard
Copied
thanks... !
🙂
Copy link to clipboard
Copied
Due to boot drive issues I had to re-install windows. I took this opportunity to run the test with only Windows and Premiere on the system. These are the latest results:
Puget Results:
Standard Overall 708
Standard Live Playback 73.2
Standard Export 68.5
GPU Score 64.2
These are much better than before.
As I re-install all my other software and drivers I will try and see which has the greatest detrimental impact to the benchmark.
Copy link to clipboard
Copied
Now that's more like it! (At least for that CPU/GPU combination.)
Copy link to clipboard
Copied
Thanks for update Richard. That's nice of you to share with others. Did you turn off indexing of drives in windows ?
Stuff like that ??
Copy link to clipboard
Copied
Here I'm after NVDIA tweak
Copy link to clipboard
Copied
That score is below those of other systems with the same CPU but lesser GPUs (I included only those results with the same major version of Premiere Pro - version 15.x). The giveaway is the relatively poor live playback score: At only 94-ish, it is well below those of other 3970X-powered PCs - even those with only an RTX 3070 or an older Titan RTX.
This tells me that there is something in your system that's bottlenecking its performance.
Copy link to clipboard
Copied
Would you share those results, I wonder what could be the bottleneck on my system and how to fix it. Thanks
Copy link to clipboard
Copied
If anyone is interested I have attached a spreadsheet of all my results. Puget have just updated the test so the newer results can't be compared to the earlier. I have noticed that my results have got worse the more my PC has become 'cluttered' with other software, perhaps the time has come for a clean install.
Copy link to clipboard
Copied
Your system is weak on the export performance, especially since my miniITX AMD Ryzen 7 3700X system consistently scored around 700 overall in the Extended score and around 740 in the Standard score, with the export score near 90.
Copy link to clipboard
Copied
@RjL190365 I thought you are serious but I definitely think that you don't know what you are talking about, pull up your test results instead!
Copy link to clipboard
Copied
Here is a result from a system with the older Titan RTX, running its RAM at only 2933 MHz:
Copy link to clipboard
Copied
And here are my scores with my ITX breadbox:
Copy link to clipboard
Copied
Copy link to clipboard
Copied
Your result is irrelevant. One cannot directly compare scores with DaVinci Resolve with those of Premiere Pro. Nor can one directly compare scores from one PugetBench benchmark to another different benchmark or even a different version of the same benchmark. And all of those different benchmarks and versions tend to make much older PCs look better than the newest PCs!