[Part 1] Measuring CPU Draw Call Performance

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

TheELF

Diamond Member
Dec 22, 2012
3,967
720
126
Slightly ahead in cine and slightly behind in other stuff,all in all about the same.
If you want you can adjust power options in windows to throttle your CPU to 75% and 50% max speed and run the bench again,that would tell what kind of scaling there is with frequency.
mpTHU98.jpg
 
Feb 17, 2017
30
2
41
Thought I'd test my good old Lynnfield as well:

CPU: Intel Core i7 860 @ 3.60Ghz
GPU: AMD Radeon HD 7950
GPU Driver: 16.12.1
OS: Windows 7 64 bit

Ships: 3000
Rocks: 6000
Draw Calls: 9021
FPS: ~9.05
 

Ansau

Member
Oct 15, 2015
40
20
81
I see 2 issues with this test and your intention to compare the data with the new Ryzen cpus.

1- Draw call capability is very sensible and dependent to cache/memory bandwidth and latencies. It is a bit pointless to compare cpus performance in this area unless they are all tested using the same speed/timings for the RAM and at stock speeds.

2- The way the settings are tweak, the framerate shows more the compute aspect of the cpu rendering the geometry rather than being a draw call focused test. Each ship contains 138 polygons, while each rock only has 28 of them.
With OP settings, the demo is throwing 679140 polygons and 9021 draw calls. If you instead max rocks slide and minimize ships slide, the demo now renders 545278 polygons and 16022 draw calls, 77% more draw calls with 24.5% less polygons.
 

MajinCry

Platinum Member
Jul 28, 2015
2,495
571
136
I see 2 issues with this test and your intention to compare the data with the new Ryzen cpus.

1- Draw call capability is very sensible and dependent to cache/memory bandwidth and latencies. It is a bit pointless to compare cpus performance in this area unless they are all tested using the same speed/timings for the RAM and at stock speeds.

2- The way the settings are tweak, the framerate shows more the compute aspect of the cpu rendering the geometry rather than being a draw call focused test. Each ship contains 138 polygons, while each rock only has 28 of them.
With OP settings, the demo is throwing 679140 polygons and 9021 draw calls. If you instead max rocks slide and minimize ships slide, the demo now renders 545278 polygons and 16022 draw calls, 77% more draw calls with 24.5% less polygons.

And if you enable instancing with the OP settings, the fps barely increases. But enable instancing with the rocks, and the fps goes >150 on my Phenom II. I'll be making a second thread with more appropriate settings.

It's intended to be more of a ballpark; if Ryzen is good with draw calls, it'll score up amongst the quad core intel CPUs. If it's not, it'll score down amongst Phenom II and Piledriver.
 

Elixer

Lifer
May 7, 2002
10,376
762
126
The draw call gap has only been widening, and AMD has not made the same leap in performance. An FX 8350 has damn near the same performance deficit with draw calls, as a 965 BE. With AMD making a big hubbub over Zen's gaming capability, they better have made the same jump as Intel, or it's dead in the water.
Not trying to be mean or anything, but, I am not sure what you are trying to show here.
It really doesn't prove anything, unless the same people who have submitted the benchmarks also do the exact same tests under Ryzen, using the same OS and everything else (which is basically impossible).
In other words, there are too many factors involved to come to any meaningful conclusion from doing tests in this manner.
 

MajinCry

Platinum Member
Jul 28, 2015
2,495
571
136
Not trying to be mean or anything, but, I am not sure what you are trying to show here.
It really doesn't prove anything, unless the same people who have submitted the benchmarks also do the exact same tests under Ryzen, using the same OS and everything else (which is basically impossible).
In other words, there are too many factors involved to come to any meaningful conclusion from doing tests in this manner.

Draw calls are only marginally affected by motherboards, RAM 'n' such. I'm looking to compare how different processors handle them, as there's clearly a discrepancy between processing prowess, and the execution of draw calls.

See my 965's scores, and the i5 2500k. The i5 is, what, 40% faster in legacy workloads? Yet it's even more superior when draw calls come into play, including when you account for differences in clocks.

I was prompted to make this test, after discussing draw calls with a guy who works with renderers for a living. http://enbseries.enbdev.com/forum/viewtopic.php?f=17&t=4869&start=10#p74912

If i remember, this was not the same in times of Pentium4, Core architecture did these improvements and each generation usually do better. According to the tests, increasing frequency of AMD cpu have very little impact on draw calls performance, while increasing cpu frequency of Intel affect it almost linear.

Also different motherboards for Intel play some role too (up to 10-20%). Not just draw calls performance, but triange count per second (in dx9 at least) limited for AMD cpu. Very stupid that AMD Athlon x2 gives about same bottleneck as tested more recent releases (for 3 or 4 years ago tested). In dx11 things should be different, but still i think the problem exists.

Even though I chose poor settings for this particular thread, it'd still give a better idea of how well Ryzen should perform in games once someone reports back their framerate when running the demo.