Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Frequently Asked Questions About NVIDIA PC (non-EFI) Graphics Cards

Status
The first post of this thread is a WikiPost and can be edited by anyone with the appropiate permissions. Your edits will be public.
Status
Not open for further replies.
tonycg1992 said:
I have a Mac Pro 3,1, and recently got a GTX 660 TI to put in it. I got it installed and working on boot up, but on games, no matter how low I put settings, the FPS stays really low. like 10-20FPS. Any suggestions? I am using DVI connection to monitor (tried both DVI slots on card) I am on el capitan now, I tried sierra and same issues. I have tried newest NVidia web drivers, and stock os driver, still same issue. NVIDIA Web Drivers give me a couple less FPS. Would bootcamp work better, or does that restrict the card?

Worth noting:

I have tried both x16, and x4 slots for the card, they both have the same FPS, both power cords plugged in for GPU, made sure GPU worked in windows PC.

I feel like something is going on though

Thanks

Specs:
2x2.8 GHz Quad-Core Intel Xeon
6GB 800MHz DDR2 FB-DIMM
NVIDIA GeForce GTX 660 Ti 2047 MB
Click to expand...

That means most likely the bottleneck is not on the GPU but something else, e.g. your CPU single core performance.

For gaming, Windows usually do much much better than OSX.
 
h9826790 said:
That means most likely the bottleneck is not on the GPU but something else, e.g. your CPU single core performance.

For gaming, Windows usually do much much better than OSX.
Click to expand...

I had a feeling, but I have seen people use this GPU/and stronger ones with this CPU, that's why I am confused. What can I do?
 
tonycg1992 said:
via bootcamp?
Click to expand...

Up to You, but bootcamp seems one of the easiest way to install Windows.

Anyway, for cMP, it's so easy to install Windows. All you need is.

1) A separate partition (strongly recommend you install Windows on a stand alone HDD / SSD)
2) Boot from the Windows installation disc (if you download the ISO from MS, you can burn it to a DVD)
3) normal Windows installation
4) Install the bootcamp driver / apps (in Windows)

The bootcamp apps in OSX is just for making partitions and help you to download the bootcamp driver installer. You don't need that.
 
h9826790 said:
Up to You, but bootcamp seems one of the easiest way to install Windows.

Anyway, for cMP, it's so easy to install Windows. All you need is.

1) A separate partition (strongly recommend you install Windows on a stand alone HDD / SSD)
2) Boot from the Windows installation disc (if you download the ISO from MS, you can burn it to a DVD)
3) normal Windows installation
4) Install the bootcamp driver / apps (in Windows)

The bootcamp apps in OSX is just for making partitions and help you to download the bootcamp driver installer. You don't need that.
Click to expand...

Won't this still bottleneck in windows though? Is there a CPU I can replace that will equal it out for a mac 3,1? Thanks so much for help.
 
tonycg1992 said:
Won't this still bottleneck in windows though? Is there a CPU I can replace that will equal it out for a mac 3,1? Thanks so much for help.
Click to expand...

If the game is very CPU limiting, yes, even in Windows, may be still not much improvement. However, usually the same games running better in Windows even with exactly the same hardware. It's not unusual that a game run 50% better in Windows. So, that means even though have the same bottleneck, most games can still perform better in Windows.
 
h9826790 said:
If the game is very CPU limiting, yes, even in Windows, may be still not much improvement. However, usually the same games running better in Windows even with exactly the same hardware. It's not unusual that a game run 50% better in Windows. So, that means even though have the same bottleneck, most games can still perform better in Windows.
Click to expand...

Well will try bootcamp to see what happens. Can I replace the CPu in 3,1 with something more on par with GPU?
 
tonycg1992 said:
Well will try bootcamp to see what happens. Can I replace the CPu in 3,1 with something more on par with GPU?
Click to expand...

Bootcamp should be a significant improvement.

Yes, you have the slowest CPU that came with 2008 Mac Pros. See the chart in my signature for compatible CPUs.
 
what games?

the 660ti is a old card so newer games will not run on high, it may also be cpu limited depending on the game (as most games only use 1-2cores (so you may have 8 but if only 2 are used then >.> the rest just chill).
the 1.1/2.1/3.1 use effectively core quad cpu's ie pre i3/i5/i7 cpus so not the fastest single core speed.

under windows you will always get better FPS much much much better FPS in windows

i was using a GTX660 till recently, worked fine for most older games but i did get a FPS boost when i moved from a macpro 3.1 to a 5.1 with the same gpu thanks to the much faster single core speed.

you shude be able to run most games from 2013 and older newer games will be harder to run
 
orph said:
what games?

the 660ti is a old card so newer games will not run on high, it may also be cpu limited depending on the game (as most games only use 1-2cores (so you may have 8 but if only 2 are used then >.> the rest just chill).
the 1.1/2.1/3.1 use effectively core quad cpu's ie pre i3/i5/i7 cpus so not the fastest single core speed.

under windows you will always get better FPS much much much better FPS in windows

i was using a GTX660 till recently, worked fine for most older games but i did get a FPS boost when i moved from a macpro 3.1 to a 5.1 with the same gpu thanks to the much faster single core speed.

you shude be able to run most games from 2013 and older newer games will be harder to run
Click to expand...

Thanks for reply. The game I am wanting to play is Guild Wars 2. In bootcamp it runs much better. Still not great when a few other players come on the screen, but doable. Would adding more ram (I have 6GB), and updating the CPU to the X5482 help my FPS at all?
 
ActionableMango said:
Those appear to be the correct (best) CPUs, but I don't have the knowledge to answer any of your other questions. I've never had a 3,1 and I don't follow used part prices.
Click to expand...

I didn't follow the eBay link, but a quick primer for the questioner:

The 3,1 supports dual 2.8, 3.0, and 3.2 gHz processor models as listed in the table. IMHO there is not much upside to upgrading the processors in a 3,1. The difference in speed is pretty minimal, nothing like the jump when upgrading a 4,1 to 5,1-compatible processors. The speed improvement is not likely to do much to close the gap with a faster GPU, if that is in fact your problem. That being said the upgrade is pretty straightforward. One other factor to consider is that the 3.2 processors are rated around 130W TDP whereas the 2.8 and 3.0 are only like 65 or 75, so you will see more heat and fans after a 3.2 upgrade.

Source: owned a 3,1 and upgraded its processors to 3.0 and then 3.2. Was not satisfied and went the 4,1->5,1 route. YMMV.
 
  • Like
Reactions: ActionableMango
flehman said:
I didn't follow the eBay link, but a quick primer for the questioner:

The 3,1 supports dual 2.8, 3.0, and 3.2 gHz processor models as listed in the table. IMHO there is not much upside to upgrading the processors in a 3,1. The difference in speed is pretty minimal, nothing like the jump when upgrading a 4,1 to 5,1-compatible processors. The speed improvement is not likely to do much to close the gap with a faster GPU, if that is in fact your problem. That being said the upgrade is pretty straightforward. One other factor to consider is that the 3.2 processors are rated around 130W TDP whereas the 2.8 and 3.0 are only like 65 or 75, so you will see more heat and fans after a 3.2 upgrade.

Source: owned a 3,1 and upgraded its processors to 3.0 and then 3.2. Was not satisfied and went the 4,1->5,1 route. YMMV.
Click to expand...
The TDP of the E5462 2.8GHz CPU is 80W whereas the TDP of the X5482 3.2GHz part is 150W. I own both a 2.8GHz 3,1 & a 3.2GHz model & see no appreciable difference in real life use use. The 15% extra performance would be apparent in benchmarks or a long render job but otherwise not. When the 4,1 was released there were complaints that the entry-level model was no faster than the 3,1 as the dual CPU 2.26GHz was slower in single stream performance than the 2.8GHz base model 3,1 & barely faster in multithreaded (due to better multithreading architecture despite lower clock speed).
 
Last edited:
  • Like
Reactions: ActionableMango
Yeah, but a 4,1/5,1 with Hexacore Xeon W3690/X5690 is significant faster than any 3,1.
 
Synchro3 said:
Yeah, but a 4,1/5,1 with Hexacore Xeon W3690/X5690 is significant faster than any 3,1.
Click to expand...
In single threaded applications it's only about 10% faster than a 3.2GHz X5482. It's faster in workloads that can take advantage of multithreading as fnot only does it have 50% more cores but also has more efficient multithreading so can do almost double the throughput. A dual processor 5,1 with 12 cores is a significant upgrade over a dual processor 3,1 with 8 cores but a single processor 5,1 with 6 cores isn't vastly faster.
 
nigelbb said:
In single threaded applications it's only about 10% faster than a 3.2GHz X5482. It's faster in workloads that can take advantage of multithreading as fnot only does it have 50% more cores but also has more efficient multithreading so can do almost double the throughput. A dual processor 5,1 with 12 cores is a significant upgrade over a dual processor 3,1 with 8 cores but a single processor 5,1 with 6 cores isn't vastly faster.
Click to expand...

W3690 GB3 single thread score is about 2800, X5482 is about 1750.

(2800-1750)/1750 = 60%

Actually a lot more than 10%.

The single thread performance is not only base on the clock speed, but improved from generation to generation.

The newer CPU can beat the old CPU in single thread performance with a lower clock speed.

In this case, The W3690 / X5690 is a newer CPU, and has higher clock speed. The result is these CPUs run significantly faster than the X5482 in both single and multi thread operation.
 
Last edited:
  • Like
Reactions: Synchro3 and pastrychef
h9826790 said:
W3690 GB3 single thread score is about 2800, X5482 is about 1750.

(2800-1750)/1750 = 60%

Actually a lot more than 10%.

The single thread performance is not only base on the clock speed, but improved from generation to generation.

The newer CPU can beat the old CPU in single thread performance with a lower clock speed.

In this case, The W3690 / X5690 is a newer CPU, and has higher clock speed. The result is these CPUs run significantly faster than the X5482 in both single and multi thread operation.
Click to expand...
Geekbench is a pretty useless benchmark. Why not look at Passmark?

X5482 (4-core 3.20GHz) Single threaded 1379
X5677 (4-core 3.47GHz) Single threaded 1518
X5690 (6-core 3.47GHz) Single threaded 1520

(1520-1379)/1379 = 10.2%

http://www.cpubenchmark.net/cpu.php?cpu=Intel+Xeon+X5482+@+3.20GHz
http://www.cpubenchmark.net/cpu.php?cpu=Intel+Xeon+X5677+@+3.47GHz
http://www.cpubenchmark.net/cpu.php?cpu=Intel+Xeon+X5690+@+3.47GHz

By way of comparison a modern CPU the 4GHz i7-6700K has single threaded performance of 2342 or 69% better than the eight year old X5482. Moore's Law hasn't applied to CPU performance for quite some years.

https://www.cpubenchmark.net/cpu.php?cpu=Intel+Core+i7-6700K+@+4.00GHz

The overall CPU mark linked to above that takes account of multithreaded performance is
X5482 (4-core 3.20GHz) 4692
X5677 (4-core 3.47GHz) 6996
X5690 (6-core 3.47GHz) 9168
7-6700K (4-core 4.0GHz) 11067
 
  • Like
Reactions: howiest
Status
Not open for further replies.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back