gpu vs cpu?

ushere wrote on 3/16/2012, 7:57 AM
ok, we've had endless discussions here about this topic, BUT...

i keep spruking vegas to an ever more knowledgeable user base who keep asking me the same question:

at what point does the gpu outperform the cpu?

wouldn't it be nice, and extremely helpful if scs, or someone would simply put together a basic table / calculator that pitted cpu vs gpu?

yes, we could all run render tests, as has been done for speed, but that still doesn't answer one of the basic questions i keep getting asked. eg. if (to use my own system as an example) if i have a i7/920 what gpu would give me faster rendering than the cpu?

or am i missing some arcane or obvious formula to find out?

Comments

megabit wrote on 3/16/2012, 8:51 AM
Well, it's very difficult to devise a simple "table" or rule, I'm afraid.

All I can say is that with my i7-2600K/GTX 580 combo, my MC AVCHD renders are 2x faster than the MC MPG2 (the same project, and equivalent output settings).

Piotr

edit: oh, and since you probably also mean the preview (not just rendering) - I would never turn my GTX 50 off now that I can get up to 8 HD tracks in MC playing full speed, at Full/Best, and with the current take full-screen on external monitor (with some color FX applied, too).

AMD TR 2990WX CPU | MSI X399 CARBON AC | 64GB RAM@XMP2933  | 2x RTX 2080Ti GPU | 4x 3TB WD Black RAID0 media drive | 3x 1TB NVMe RAID0 cache drive | SSD SATA system drive | AX1600i PSU | Decklink 12G Extreme | Samsung UHD reference monitor (calibrated)

Dach wrote on 3/16/2012, 10:36 AM
I asked this same question just the other day and am in process of doing a couple comparisons.

Using the test files that Sony provided I have done three tests. Rendering to a 1080 MP4 file.

System 1: Phenom II 945e @3000 MHz, 16 GB RAM GPU Turned Off (due to irrelevant graphics card. Render Time: 09:16

System 2: Phenom II 955e @3200 MHz, 8 GB RAMGeforce 550Ti Turned On Render Time: 04:24

I was really impressed the improved peformance. I then turned off GPU acceleration on System 2 and received a render time of approximately 04:30 (don't have my notes in front of me).

While I still want to redo my tests, this raises the question. What was the contributing factor.

System 1 does run an older AMD 785 Chipset, wich I know is a contributing factor to its lower performance. I will be conducting a test on another machine I have, which is an Athalon II which benchmarks better with the AMD 880 Chipset.

Chad

megabit wrote on 3/16/2012, 10:44 AM
Let's get more specific here:

The GPU can be turned off/on in 2 different places:

1. VP 11 Preferences->Video

2. some render templates (Sony and MC AVCHD)

Now: where did you switch your GPU off? 'Cause I reckon point 1 applies mainly to the preview, while point 2 - to the rendering acceleration...

Piotr

AMD TR 2990WX CPU | MSI X399 CARBON AC | 64GB RAM@XMP2933  | 2x RTX 2080Ti GPU | 4x 3TB WD Black RAID0 media drive | 3x 1TB NVMe RAID0 cache drive | SSD SATA system drive | AX1600i PSU | Decklink 12G Extreme | Samsung UHD reference monitor (calibrated)

craftech wrote on 3/16/2012, 10:46 AM
I believe this may be at the core of all the problems with Vegas 11 so while the CPU dependent versions of Vegas may be a bit slower, I believe they were more reliable and that to me is more important than speed.

John