Official Statement from Otoy towards GTX 980/970
Forum rules
Please add your OS and Hardware Configuration in your signature, it makes it easier for us to help you analyze problems. Example: Win 7 64 | Geforce GTX680 | i7 3770 | 16GB
Please add your OS and Hardware Configuration in your signature, it makes it easier for us to help you analyze problems. Example: Win 7 64 | Geforce GTX680 | i7 3770 | 16GB
this forums is full of very interesting people =) hope OTOY will make/sponsor some future events to meet You all =)
- Tutor
- Posts: 531
- Joined: Tue Nov 20, 2012 2:57 pm
- Location: Suburb of Birmingham, AL - Home of the Birmingham Civil Rights Institute
Like the EVGA GTX 980 Classified that I mentioned in one of my earlier posts here, the 4G Galax(y) HOF [ http://videocardz.com/53918/galax-gefor ... f-pictured ] may also be a Maxwell GPU that shines in Octane rendering. It too has 2x 8-pin power connectors and a very high TDP (in fact, a little higher than that of the 4G EVGA GTX 980 Classified). Also, it appears to have better coolers.
Because I have 180+ GPU processers in 16 tweaked/multiOS systems - Character limit prevents detailed stats.
"bit too much of that bling" =DDD..but I'd be interested to see how it swims (I mean performs under waterblock =)Tutor wrote:Like the EVGA GTX 980 Classified that I mentioned in one of my earlier posts here, the 4G Galax(y) HOF [ http://videocardz.com/53918/galax-gefor ... f-pictured ] may also be a Maxwell GPU that shines in Octane rendering. It too has 2x 8-pin power connectors and a very high TDP (in fact, a little higher than that of the 4G EVGA GTX 980 Classified). Also, it appears to have better coolers.
& that white PCB is plain sexy =DDD..
That sounds very much like mine, which is a MSI GTX 980 tested here.Tutor wrote:Like the EVGA GTX 980 Classified that I mentioned in one of my earlier posts here, the 4G Galax(y) HOF [ http://videocardz.com/53918/galax-gefor ... f-pictured ] may also be a Maxwell GPU that shines in Octane rendering. It too has 2x 8-pin power connectors and a very high TDP (in fact, a little higher than that of the 4G EVGA GTX 980 Classified). Also, it appears to have better coolers.
http://www.overclockersclub.com/reviews ... gaming_4g/
It has two 8-pin power connects and they got it overclocked to 1525 MHz on air. I'm using Linux so I don't do much with overclocking, but according to Nvidia's X settings tool mine is running at around 1400 MHz but I don't know if that is the core or boost clock. Memory is at 7000 MHz. I haven't noticed too much difference in performance from the other 980's people have tested unfortunately.
Linux Mint 21.3 x64 | Nvidia GTX 980 4GB (displays) RTX 2070 8GB| Intel I7 5820K 3.8 Ghz | 32Gb Memory | Nvidia Driver 535.171
- Tutor
- Posts: 531
- Joined: Tue Nov 20, 2012 2:57 pm
- Location: Suburb of Birmingham, AL - Home of the Birmingham Civil Rights Institute
Hello Grimm and pBarrelas,
Since one of you has a dual 8-pin powered GTX 980 and the other appears to have a dual 6-pin powered GTX 980, then maybe we can learn whether my hypothesis about throttling has legitimacy if:
1) both of your 980's GPU and memory speeds can be set to match each other,
2) you both use the same/latest CUDA and Nvidia Driver versions,
3) each of you runs, three times, the same version of the Octane Demo scatter.ocs test (but it would be better if each of you runs that test, three times, using Octane Demo v1.2 and v1.2 of scatter.ocs and then, three times, using Octane Demo v2.06 and v2.06 of scatter.ocs), and
4) each of you tell me your GPUs' completion times and ms/s for each pass.
Thanks for sharing your situation. I suspect that the 1400 MHz may be the boost clocking. When you say, "I haven't noticed too much difference in performance from the other 980's people have tested unfortunately," are you referring to it's performance in Octane and, if so, using what metric. For example, are you using an Octane demo test scene and if so which one and what's the version no.? Are you using render completion time in secs and/or s/px? Also, do you have access to Octane V.1.2 and the V1.2 demo (this is to see whether the Octane version makes a significant difference) ? What version of CUDA and Nvidia drivers are you using? Can you underclock your GPU?
Tutor
pBarrelas,
I suspect that an EVGA GTX 980 8gb Classified will not perform any better in Octane than an EVGA GTX 980 4gb Classified, except that the former will not as easily choke when rendering UltraHighDefinition, high density scenes (like 4K or above). As to whether and why an EVGA GTX 980 Classified, similarly configured and using the same version of CUDA and Nvidia drivers, might out-perform the card that you purchased, see my earlier response (just above this one) to Grimm.
Tutor
Since one of you has a dual 8-pin powered GTX 980 and the other appears to have a dual 6-pin powered GTX 980, then maybe we can learn whether my hypothesis about throttling has legitimacy if:
1) both of your 980's GPU and memory speeds can be set to match each other,
2) you both use the same/latest CUDA and Nvidia Driver versions,
3) each of you runs, three times, the same version of the Octane Demo scatter.ocs test (but it would be better if each of you runs that test, three times, using Octane Demo v1.2 and v1.2 of scatter.ocs and then, three times, using Octane Demo v2.06 and v2.06 of scatter.ocs), and
4) each of you tell me your GPUs' completion times and ms/s for each pass.
Grimm,grimm wrote:That sounds very much like mine, which is a MSI GTX 980 tested here.Tutor wrote:Like the EVGA GTX 980 Classified that I mentioned in one of my earlier posts here, the 4G Galax(y) HOF [ http://videocardz.com/53918/galax-gefor ... f-pictured ] may also be a Maxwell GPU that shines in Octane rendering. It too has 2x 8-pin power connectors and a very high TDP (in fact, a little higher than that of the 4G EVGA GTX 980 Classified). Also, it appears to have better coolers.
http://www.overclockersclub.com/reviews ... gaming_4g/
It has two 8-pin power connects and they got it overclocked to 1525 MHz on air. I'm using Linux so I don't do much with overclocking, but according to Nvidia's X settings tool mine is running at around 1400 MHz but I don't know if that is the core or boost clock. Memory is at 7000 MHz. I haven't noticed too much difference in performance from the other 980's people have tested unfortunately.
Thanks for sharing your situation. I suspect that the 1400 MHz may be the boost clocking. When you say, "I haven't noticed too much difference in performance from the other 980's people have tested unfortunately," are you referring to it's performance in Octane and, if so, using what metric. For example, are you using an Octane demo test scene and if so which one and what's the version no.? Are you using render completion time in secs and/or s/px? Also, do you have access to Octane V.1.2 and the V1.2 demo (this is to see whether the Octane version makes a significant difference) ? What version of CUDA and Nvidia drivers are you using? Can you underclock your GPU?
Tutor
pBarrelas wrote:Yeah, I'm very curious to see how an EVGA GTX 980 8gb Classified, performs in Octane, as I have an Asus GTX 980 4gb purchased before starting to use Octane.
pBarrelas,
I suspect that an EVGA GTX 980 8gb Classified will not perform any better in Octane than an EVGA GTX 980 4gb Classified, except that the former will not as easily choke when rendering UltraHighDefinition, high density scenes (like 4K or above). As to whether and why an EVGA GTX 980 Classified, similarly configured and using the same version of CUDA and Nvidia drivers, might out-perform the card that you purchased, see my earlier response (just above this one) to Grimm.
Tutor
Last edited by Tutor on Thu Nov 06, 2014 5:54 pm, edited 2 times in total.
Because I have 180+ GPU processers in 16 tweaked/multiOS systems - Character limit prevents detailed stats.
- Tutor
- Posts: 531
- Joined: Tue Nov 20, 2012 2:57 pm
- Location: Suburb of Birmingham, AL - Home of the Birmingham Civil Rights Institute
Glimpse,glimpse wrote:"bit too much of that bling" =DDD..but I'd be interested to see how it swims (I mean performs under waterblock =)Tutor wrote:Like the EVGA GTX 980 Classified that I mentioned in one of my earlier posts here, the 4G Galax(y) HOF [ http://videocardz.com/53918/galax-gefor ... f-pictured ] may also be a Maxwell GPU that shines in Octane rendering. It too has 2x 8-pin power connectors and a very high TDP (in fact, a little higher than that of the 4G EVGA GTX 980 Classified). Also, it appears to have better coolers.
& that white PCB is plain sexy =DDD..
If you know about water cooling GPUs, I need your help/advice because that's what I plan to do to four or five Titans and an EVGA 8G GTX 980 Classified.
Tutor
Because I have 180+ GPU processers in 16 tweaked/multiOS systems - Character limit prevents detailed stats.
I'll run the test but 1st I got a couple of questions, since I'm pretty new to Octane:
1-Where can I find Octane Demo scatter.ocs?
2-Why running theses tests using 1.2 and 2.06 demos and not the licensed versions?
I'm sure these are stupid questions but I've been using Octane from last Saturday, so a lot of this stuff is unfamiliar to me!
1-Where can I find Octane Demo scatter.ocs?
2-Why running theses tests using 1.2 and 2.06 demos and not the licensed versions?
I'm sure these are stupid questions but I've been using Octane from last Saturday, so a lot of this stuff is unfamiliar to me!
Win 8.1 Pro 64-bit| i7-4930k 3.4GHz | 32 Gb RAM | GTX 980 4Gb| Driver Version 344.75
Hi Tutor,
I have similar questions as pBarrelas. The tests I have been running have been for both Octane and Blender Cycles.
I tried running Octane 1.2 but it doesn't recognize the 980 so I used 1.55 instead. This is using the benchmark trench scene completed to 1000 s/px.
DL - 19.63 Ms/sec, 1000 samp/px - 31 sec.
PT - 5.96 Ms/sec, 1000 samp/px - 1 min. 41 sec.
PMC - 4.4 Ms/sec, 1000 samp/px - 2 min. 18 sec. (this has the alpha shadows on, turning them off errors out Octane and I had to restart the program)
These are the times for version 2.12.1:
DL - 15.07 Ms/sec, 1000 samp/px - 40 sec.
PT - 4.16 Ms/sec, 1000 samp/px - 2 min. 28 sec.
PMC - 3.62 Ms/sec, 1000 samp/px - 2 min. 46 sec.
These numbers should change once we get to use the next version of Octane. Marcus has already been able to improved the speed of the PT kernel to within 10 to 20 percent of the speed of a Titan Black.
I have similar questions as pBarrelas. The tests I have been running have been for both Octane and Blender Cycles.
I tried running Octane 1.2 but it doesn't recognize the 980 so I used 1.55 instead. This is using the benchmark trench scene completed to 1000 s/px.
DL - 19.63 Ms/sec, 1000 samp/px - 31 sec.
PT - 5.96 Ms/sec, 1000 samp/px - 1 min. 41 sec.
PMC - 4.4 Ms/sec, 1000 samp/px - 2 min. 18 sec. (this has the alpha shadows on, turning them off errors out Octane and I had to restart the program)
These are the times for version 2.12.1:
DL - 15.07 Ms/sec, 1000 samp/px - 40 sec.
PT - 4.16 Ms/sec, 1000 samp/px - 2 min. 28 sec.
PMC - 3.62 Ms/sec, 1000 samp/px - 2 min. 46 sec.
These numbers should change once we get to use the next version of Octane. Marcus has already been able to improved the speed of the PT kernel to within 10 to 20 percent of the speed of a Titan Black.
Linux Mint 21.3 x64 | Nvidia GTX 980 4GB (displays) RTX 2070 8GB| Intel I7 5820K 3.8 Ghz | 32Gb Memory | Nvidia Driver 535.171
pBarrelas wrote:I'll run the test but 1st I got a couple of questions, since I'm pretty new to Octane:
1-Where can I find Octane Demo scatter.ocs?
2-Why running theses tests using 1.2 and 2.06 demos and not the licensed versions?
I'm sure these are stupid questions but I've been using Octane from last Saturday, so a lot of this stuff is unfamiliar to me!
Test scene
http://render.otoy.com/forum/viewtopic. ... +your+rigs
http://keithlea.com/in/OctaneBenchmark.zip
For testing use 2.12.1
Don't modify any view in the viewport (don't move, rotate, zoom)- it affects the speed. This is IMPORTANT.
Please fill in the scores for the following settings for 2.12.1 - this will give us a general idea of how your card performs under different settings.
2.12 (these are scores for my rig)
DL 171.45 ms/s
DL 190.91 ms/s (coherent mode)
PT 32.28 ms/s (path term power 0.3)
PT 34.78 ms/s (path term power 0.3, coherent mode)
PT 60.27 ms/s (path term power 1, coherent mode)
PMC 24.48 ms/s
PMC 21.45 ms/s (alpha shadows on)
You may also be willing to give us your gpu and mem. clocks - msi afterburner gives them all.
3090, Titan, Quadro, Xeon Scalable Supermicro, 768GB RAM; Sketchup Pro, Classical Architecture.
Custom alloy powder coated laser cut cases, Autodesk metal-sheet 3D modelling.
build-log http://render.otoy.com/forum/viewtopic.php?f=9&t=42540
Custom alloy powder coated laser cut cases, Autodesk metal-sheet 3D modelling.
build-log http://render.otoy.com/forum/viewtopic.php?f=9&t=42540