OTOY Forums

Hi All

Does anyone know when or if the Octane Render engine will use CPU if there is no compatible GPU available?

think it was advertised in v3, so let's way to see, maybe next year =) only OTOY knows for sure, but I guess there are more features that community would prefer to see & anyway on CPUs it would be painfully slow.. What are You using? (that have a need for CPU rendering)

Sometimes when I go away and stay with friends all I have is my old macbook pro with windows 7 on it and it will not run Octane.
It would be great to be able to render things while I am away sometimes.

glimpse wrote:...........anyway on CPUs it would be painfully slow..

Well that depends on their optimizations and implementation. It doesn't need to be painfully slow(er). Look at Corona render engine how fast is it on CPU.

Also i'm occasionally using my rig for Bitcoin (actually Altcoins) mining and in my case single CPU Xeon E5-2698v4 is slightly faster than GTX 1080 (550 Hashes/second vs 535 H/s for GTX 1080 GPU on CCminer WITH CUDA 8.0 support)
Titan X(M) on cuda 7.5 and 8.0 cuda enabled ccminer gives only 150-160 H/s so this thing is heavily optimized for cuda64_80.dll.

I think Octane could/should have lot more CUD a8.0 optimizations for Pascal GPUs 'coz from example above there is huge benefit with all optimizations while in Octane my TitanX(M) is same speed as Pascal 1080 and in CCminer mining Pascal GPUs smashesh Maxwell so much like 3x since optimized cuda 8.0 miner is on market.

Lewis wrote: Well that depends on their optimizations and implementation. It doesn't need to be painfully slow(er). Look at Corona render engine how fast is it on CPU.

I'd rather see OTOY working on improving features & only then widening compatibility. Quality over quantity in other words.

Also i'm occasionally using my rig for Bitcoin (actually Altcoins) mining and in my case single CPU Xeon E5-2698v4 is slightly faster than GTX 1080 (550 Hashes/second vs 535 H/s for GTX 1080 GPU on CCminer WITH CUDA 8.0 support)
Titan X(M) on cuda 7.5 and 8.0 cuda enabled ccminer gives only 150-160 H/s so this thing is heavily optimized for cuda64_80.dll.

There are different workloads some of them would fly on CPU & lag on GPU while running other would be opposite story.
It's like, saying..that expensive tractor is faster on the muddy field compared to Ducati Panigale - story would be different if they would be on track. The same applies here as well.

By the way, how fast it's possible to break even with mining at this time? Was thinking about putting small AMD based rig in the background to do some mining till we get support in OctaneRender for those cards =)

Lewis wrote:I think Octane could/should have lot more CUD a8.0 optimizations for Pascal GPUs 'coz from example above there is huge benefit with all optimizations while in Octane my TitanX(M) is same speed as Pascal 1080 and in CCminer mining Pascal GPUs smashesh Maxwell so much like 3x since optimized cuda 8.0 miner is on market.

now that's interesting. Actually Jules posted on OctaneRender facebook group that OTOY made progress optimizing engine with help of nvidia & new release will come ;) fingers crossed we are going to see 200 on TitanX (Pascal) would love to grab that one or two ;)