• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

[Part 2] 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.
.vodka said:
How come the G3258 rig is posting 30FPS while the 6700k rig is at the 25 FPS mark? They're using different drivers, and it's Windows 7 vs Windows 10, but still.
Click to expand...
I mean, that would cover it all. IPC and cache do not come into play here since it is basically a cycle of interrupts.
 
.vodka said:
It doesn't, 6700k is 4GHz base clock, 4.2GHz single threaded turbo and is using a GTX1080, so it's nvidia vs nvidia.

Assuming this is a single threaded test, 100Mhz advantage for the G3258 doesn't even cover the IPC gap between full Skylake 6700k and cut down (3MB L3$) Haswell G3258.

How come the G3258 rig is posting 30FPS while the 6700k rig is at the 25 FPS mark? They're using different drivers, and it's Windows 7 vs Windows 10, but still.
Click to expand...

I'd say that the OS makes up for that. We saw in the previous test that Win 10 gave two extra fps over Win XP. Wouldn't be surprised if Win 10 gives 1 fps over Win 7 in that test.

Now that we have more optimal driver settings, that difference would be more pronounced. Couple that with a change in GPU brand to NVidia, this is as best-case a test can be for Green team. Also keep in mind the almost linear relation between "clocks:draw calls" on Intel CPUs.
 
.vodka said:
It doesn't, 6700k is 4GHz base clock, 4.2GHz single threaded turbo and is using a GTX1080, so it's nvidia vs nvidia.

Assuming this is a single threaded test, 100Mhz advantage for the G3258 doesn't even cover the IPC gap between full Skylake 6700k and cut down (3MB L3$) Haswell G3258.

How come the G3258 rig is posting 30FPS while the 6700k rig is at the 25 FPS mark? They're using different drivers, and it's Windows 7 vs Windows 10, but still.
Click to expand...

Yeah, idk why it's performing that well.

Cool I guess. 2 cores still kinda sucks.
 
dogen1 said:
Yeah, idk why it's performing that well.

Cool I guess. 2 cores still kinda sucks.
Click to expand...

'Tis the driver. Outside of this optimal scenario, that has no meaningful lighting, shadows, shaders, materials and all that good stuff, it wouldn't show such a lead.

You got an AMD GPU lying around you can chuck in?
 
MajinCry said:
'Tis the driver. Outside of this optimal scenario, that has no meaningful lighting, shadows, shaders, materials and all that good stuff, it wouldn't show such a lead.

You got an AMD GPU lying around you can chuck in?
Click to expand...

I have an old 4850 that probably doesn't work anymore lol.

And, with the latest driver and my browser closed(every little bit helps i guess) I hit pretty consistent 32.5 now.

Little beast of a cpu :openmouth:

edit - actually it seems to vary a bit.

made a screenshot if anyone cares to see

20RjqqC.png
 
Last edited:
CPU: i7 4771
GPU: XFX R9 Nano
GPU Driver: 17.1.2 with settings specified in OP
OS: Windows 10 x64

Ships: 1
Rocks: 16000
Draw Calls: 16022
FPS: 17.78-18.67

csgOFqX.jpg
 
Updated graph with TheStilt's R7 1800x results. Of course, this is just one result in a synthetic benchmark, but it doesn't look good.
 
CPU: Ryzen 1800x @ 4.0ghz (SMT OFF)
GPU: Radeon 290 (stock)
GPU Driver: 17.2.1
RAM: 32GB DDR4 2666 14-14-14-34 (rated for 3200 but won't run above 2666 no matter what)
MOBO: Asus Crosshair VI Hero
OS: Windows 10 Home 64-bit

Ships: 1
Rocks: 16000
Draw Calls: 16022
FPS: 16.53-17.13

K0qiEzy.jpg

Take any Ryzen results with a bag of salt atm. I ran this once and averaged 17fps, but throughout my numerous reboots I began only getting 14 fps. Then as I was writing this up, my test gave me 17fps again. What the frick?

I've been messing around with my setup since about 5pm ET today, and the BIOS for this Asus Crosshair mobo is a complete mess. When I first booted it up the CPU voltage was at 1.5+ pretty much constantly. Now it's mysteriously gone down, but I'm still getting strange results. Even at stock speeds it's pushing 1.4v when set to 'auto'. I've been in and out changing all sorts of settings and nothing ever seems to make much sense. Some changes don't actually take effect, some seem to cause other settings to mysteriously change, etc.

**edit** Added screenshot of my Radeon settings:

7UDpKiC.jpg
 
Last edited:
CPU: G3258 @ 4.4GHz
GPU: R9 270
GPU Driver: 16.11.5
OS: Win 7 64bit

Ships: 1
Rocks: 16000
Draw Calls: 16022
FPS: 28.02-28.35
Capture_2017_03_06_19_47_26_866.png
 
tomatone said:
Duno this is my result with 3000 ships/6000 rocks =15.97-16.28FPS
Click to expand...

The problem with any settings other than max rocks, is that the application thread takes up damn near all of the CPU time, which is why Part 1 of this test was a failure.

It's just bizarre man. Your Haswell Pentium with an AMD GPU scores up with the fully fledged i7's paired with an NVidia GPU (NV's driver being "optimized" for such synthetics).
 
CPU: C2D E8400 E0 @ 3GHz (stock)
GPU: ATI Radeon HD 4850 512MB
GPU Driver: 8.970.100.1100 WHQL (Catalyst 13.9)
OS: Win 7 x64

Ships: 1
Rocks: 16000
Draw Calls: 16022
FPS: 11.84-11.91 window focused, 9.59-9.61 not focused, 12.12-12.15 fullscreen

GPU core/mem clocks have no effect. Texture Preference and MipMap detail level settings at ati tray tools have no effect (I don't think I have anything closer than that to your driver screen).
 
MajinCry said:
It's just bizarre man. Your Haswell Pentium with an AMD GPU scores up with the fully fledged i7's paired with an NVidia GPU (NV's driver being "optimized" for such synthetics).
Click to expand...

Haswell Pentium is ridiculously OP in this benchmark so far.

There's some sorcery going on in here.
 
nevbie said:
CPU: C2D E8400 E0 @ 3GHz (stock)
GPU: ATI Radeon HD 4850 512MB
GPU Driver: 8.970.100.1100 WHQL (Catalyst 13.9)
OS: Win 7 x64

Ships: 1
Rocks: 16000
Draw Calls: 16022
FPS: 11.84-11.91 window focused, 9.59-9.61 not focused, 12.12-12.15 fullscreen

GPU core/mem clocks have no effect. Texture Preference and MipMap detail level settings at ati tray tools have no effect (I don't think I have anything closer than that to your driver screen).
Click to expand...

Sweet. Updated the results. All we need now is a highly clocked C2D, and we'd be able to tell if higher clocks put Core 2 past Phenom II, in draw calls.


.vodka said:
Haswell Pentium is ridiculously OP in this benchmark so far.

There's some sorcery going on in here.
Click to expand...

Guess if you're going to be playing modded Morrowind or Oblivion, an OC'd Haswell Pentium is the way forward, for some crazy reason.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top