Just curious, which 5,1 benchmark was faster? I looked back through the thread, maybe I just missed it.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Unigene Superposition Benchmark
- Thread starter Prince134
- Start date
- Sort by reaction score
Why is your i7-7700K with GTX 1080 Ti 14% faster than my PC with the exact same configuration? https://forums.macrumors.com/threads/unigene-superposition-benchmark.2043107/page-2#post-24613504
Update: Ok I see your GPU temperature is much lower. I guess water cooling?
So the cooling makes more difference than the CPU Generation. And the cooler of the GTX 1080 Ti Founders Edition is miserable.
Without doubt the Mac Pro 4,1/5,1 is still a great machine.
Last edited:
I have the 7700K at 5.0GHz 24/7 OC and the Ti is at 2050-2060MHz Ti RAM at 5800 and computer system RAM is at 3600Mhz.
It is all on water cooling. The Ti will idle around 28-30C, CPU 34-36 and will hold this temp on essentially silent fans during normal use such as watching youtube, internet browsing, etc even with gaming the temps will stay in the 38-42C range on the GPU and 60's for the CPU.
The only time this card will see high 40's will be during specific benchmarking tools.
I also had this machine running on Sierra up until a week ago and it did very well.
I'm thinking of buying a bigger PC case and set all under water.
Which case and water cooling system do you use?
I actually just did a Mid Tower, I am not a fan of the huge full tower cases.I'm thinking of buying a bigger PC case
and set all under water.
Which case and water cooling system do you use?
This is my first ever hard tube custom loop setup. Prior to this is just used the all-in-one coolers like from Corsair.
Even on the GPU. My sons PC has a water cooled 1070 with a Corsair H55. That card stays at 43-45C during benchmarking and OC'd to 2126MHz. Prior to the 1070 I had SLI 980Ti's both on Corsair H55's.
Currently I am using all EKWB parts in a Corsair 570X Crystal tempered glass case.
It's a very 'flashy' completely different build than I have ever done previously.
I also jus re-ran the benchmarks and was surprised to see a slight increase across the board.
Last edited:
You should have told us your Unigene scores were overclocked.
Your CPU is 47% higher clocked, GPU higher by 30%, but only get 10% higher score in 1080P Extreme? what was missing?
Last edited:
Yep. That's totally not very honest IMO.
Looks like the 5,1 kicks ass in gaming, even compared to a 7700k.
Looks like the 5,1 kicks ass in gaming, even compared to a 7700k.
Prince, it seems that you have a Titan and a 1080Ti.
Can you run the VR Maximum and VR Optimum on the 1080 Ti?
I see that you did it on the Titan, but not the new Pascal card.
Can you run the VR Maximum and VR Optimum on the 1080 Ti?
I see that you did it on the Titan, but not the new Pascal card.
Thanks Prince.
What's a good GTX card for a 5,1 Mac Pro with a 3.46/6-core for gaming at 1440p that wants to start investing in VR?
What's a good GTX card for a 5,1 Mac Pro with a 3.46/6-core for gaming at 1440p that wants to start investing in VR?
GTX 1080 is the best I think.
Can't beat the price:
https://slickdeals.net/f/10202272-g...080-ti-gaming-oc-11-gb-634-99-via-newegg-ebay
VR Maximum:
VR optimum:
Can't beat the price:
https://slickdeals.net/f/10202272-g...080-ti-gaming-oc-11-gb-634-99-via-newegg-ebay
VR Maximum:
VR optimum:
Last edited:
Good stuff. Thanks!
It appears that a 1080ti may be significantly weaker in a Mac Pro 5,1 from my research. A modern (or semi recent) i7 would push the GTX 1080ti around 9500 or above for the VR maximum test.
Of course that should be vetted. I may be wrong!
It still appears that the Mac Pro would do current VR just fine.
I wonder the difference in your score that a GTX 1070 would make, or a regular GTX 1080.
I wonder if there is a degradation point with our old test beds...
[doublepost=1496295281][/doublepost]Syncro, do you mind running the VR tests with your 1080ti? That would tell a damn good story for us all in how the modern 7700k compares to a westmere in VR performance. Would love to see it.
It looks like Prince's 3.46 Westy is putting up some impressive numbers and has overtook some of the 2d numbers produced by your 7700k. Not sure how that's possible but....
It appears that a 1080ti may be significantly weaker in a Mac Pro 5,1 from my research. A modern (or semi recent) i7 would push the GTX 1080ti around 9500 or above for the VR maximum test.
Of course that should be vetted. I may be wrong!
It still appears that the Mac Pro would do current VR just fine.
I wonder the difference in your score that a GTX 1070 would make, or a regular GTX 1080.
I wonder if there is a degradation point with our old test beds...
[doublepost=1496295281][/doublepost]Syncro, do you mind running the VR tests with your 1080ti? That would tell a damn good story for us all in how the modern 7700k compares to a westmere in VR performance. Would love to see it.
It looks like Prince's 3.46 Westy is putting up some impressive numbers and has overtook some of the 2d numbers produced by your 7700k. Not sure how that's possible but....
Last edited:
Note that Prince134 uses vanilla gpu, which in mac pro on bootcamp runs on pcie 1.1.
When efi-flashed by macvidcards the gpu is "unlocked" on bootcamp and runs on pcie 2.0, which is faster and should bring those benchmark points a little bit higher IMO.
That's why I asked Syncro to run. We haven't gotten to the bottom of this. There shouldn't be a reason why your machine is faster than synchros.
Well I apologize. I don't always have time for a large post at work.
I posted the actual Unigine benchmark information on their website. I noticed at the time it did indeed show the GPU's overclocked frequencies on all benchmarks. I'm not sure why the CPU doesn't reflect it's OC. I was paying more attention to the GPU info listed on the benchmark.
I think we need to see what a STOCK i7 7700 does with a STOCK GTX 1080Ti in the VR benchmarks to see how well a Mac Pro 5,1 does in VR and to understand more about how much, or even "if" it's CPU-bound.
Posting 60% overclock'd speeds from a PC doesn't tell us anything. Any chance you could re-run the benches with your gear at stock speeds and compare it with Prince's latest posts?
Posting 60% overclock'd speeds from a PC doesn't tell us anything. Any chance you could re-run the benches with your gear at stock speeds and compare it with Prince's latest posts?
I sure can.
I'll remote in and let benchmarks run and post once they are completed.
Kick-ass! If you could, I wouldn't mind seeing the 2D benches as well. But only if you have time!
This is exciting!
This is exciting!
Sorry, every time I want to run a VR test the display turns black a few seconds and I'm back on the main menu of Unigine Superposition.... Nothing happens. However, I'm using an Oculus Rift without problems, with the Mac Pro and with the i7-PC.
My Mini-ITX case seems too small for an i7-7700K and a GTX 1080 Ti, here an old pic then with a GTX 980 Ti: https://forums.macrumors.com/threads/the-hackintosh-thread.1900326/page-20#post-24274152
I opened the Mini-ITX case in order to give more air flow to the GTX 1080 Ti Founders Edition and ran all benchmarks again. Nevertheless every time the card reaches 84° C it throttles down to 1500-1600 MHz...
Results now a bit better than with closed Mini-ITX case.
i7-7700K (no OC) with GTX 1080 Ti FE (no OC):
[doublepost=1496350341][/doublepost]
Update: More results with other GTX 1080 Ti cards here: https://forums.evga.com/Unigine-Superposition-Benchmarks-for-GTX-10-series-m2646934.aspx The water cooled cards (all cards that do not reach 84-85°) are superior.
Last edited:
Awesome, thanks! Looks like your scores have gone up. Seems like the 5,1 is still inching ahead. Probably as you say though, there is some throttling going on from being in a mini ATX case.
Really appreciate the data!
Really appreciate the data!
Yes, the air flow from the CPU goes straight to the backplate of the graphics card in the small ITX case. Not an advantage.
But the original cooler of the GTX 1080 Ti Founders Edition is miserable too.
I think an Alphacool Eiswolf 120 GPX Pro could be a good solution:
https://www.alphacool.com/shop/-new...eforce-gtx-titan-x-pascal-/-1080-ti-m02-black
But the original cooler of the GTX 1080 Ti Founders Edition is miserable too.
I think an Alphacool Eiswolf 120 GPX Pro could be a good solution:
https://www.alphacool.com/shop/-new...eforce-gtx-titan-x-pascal-/-1080-ti-m02-black
Last edited:
Some findings. Turns out Nvidia has tuned the 1080 Ti Founders Edition by default to keep max fan speeds below about 50 percent (2400 rpm), resulting in 84°C GPU temperatures and thermal throttling. Once the GTX 1080 Ti FE reaches 84°C it throttles down the GPU.
Superposition 1080p Extreme, throttling from 1885 MHz to 1630 MHz:
Superposition 4K optimized, throttling from 1809 MHz to 1530 MHz:
Solution: Set the GPU fan speed with the EVGA Precision tool to 70% (3360 rpm), which is certainly audible, to avoid thermal throttling. 80% and more is very loud.
Setup: CPU i7-7700k (no OC) with ASUS GTX 1080 Ti Founders Edition (no OC, only fan speed set to 70%), RAM 2400 MHz, ASRock Z270 Mini ITX, Air cooling.
New results:
The i7-7700K (no OC) is almost 1% faster than the X5690... Superposition is GPU bound, CPU does not matter.
Superposition 1080p Extreme, throttling from 1885 MHz to 1630 MHz:
Superposition 4K optimized, throttling from 1809 MHz to 1530 MHz:
Solution: Set the GPU fan speed with the EVGA Precision tool to 70% (3360 rpm), which is certainly audible, to avoid thermal throttling. 80% and more is very loud.
Setup: CPU i7-7700k (no OC) with ASUS GTX 1080 Ti Founders Edition (no OC, only fan speed set to 70%), RAM 2400 MHz, ASRock Z270 Mini ITX, Air cooling.
New results:
The i7-7700K (no OC) is almost 1% faster than the X5690...
Superposition is GPU bound, CPU does not matter.
Last edited: