T_Flight

CraptacularOne


And as for 400 series owners not having any experience with 500 series cards, that's really just one of the dumbest things I've heard in a while. Both cards use the same architecture (Fermi) and what applies to one, will apply to the other. If both are used with the same driver, both will show roughly the same percentage gains and trends between them. Of course the review above isn't using the same driver, yet still yields strikingly similar gains in the games tested in regards to performance and scaling. What does that tell us?

Horribly Ignorant huh? The above is the epitome of Ignorant. It shows in perfect clarity why those without experience with a card in this forum needs to keep quiet about things they know nothing about. 4xx series have WILDLY different issues than 5xx series cards, an NO (HELL NO) what applies to one doesn't apply to the other. That's precisely the crap that is littering this forum up and is causing people greif here. The mods should lock these forums to people who actuall own the hardware of the forum they are posting in.

That review tells you NOTHING about 5xx series cards with a dedicated PhysX card as I told you in my post before it. This is NOT the 4xx series forum, nor does anything related to them belong here at all. Spin that however you like, but the fact is you're wrong, as the OP has shown, and as I have shown you along with others WHO ACTUALLY OWN these cards.

On another note, 4xx series have drivers that are not even compatible with 5xx until the recently released 266.xx series so again, you're comparing apples to oranges.

The only similarity between these cards is the Fermi name. 5xx has a different board layout, different ram, GF110 GPU verses GF100, wider RAM bus, fully unlocked shaders, different cooling, completely different OCP and OVP scheme, and they have completely different issues like 120 Hz/Multi Monitor setups not clocking down by design just to name one of many.

There are two separate forums for a reason...use them.

Well okay...so you're off on another rant. Care so show us some hard data that disagrees with what was posted or just blurt out another wall of text with NOTHING backing it again? I'll have a pair of 580's shortly and I will post benches and actual data. You're not a mod and you've no business telling anyone where and how to post. Leave that up to the mods.

CraptacularOne

*sigh*..... is there no end to the ignorance here?

Normally I wouldn't do this. But I am really just getting fed up with people who know nothing and then upon being shown otherwise, refuse to accept that they are wrong.

This card is for a client's PC that arrived this morning for a build I'm doing for him. I'm sure he won't mind if I pop it in for a quick "test run".



So what do you say we fire up a quick benchmark? Lets see if there truly is no difference with a dedicated PhysX card or not as the less informed want to whine and cry about.

Mafia II GTX570 no dedicated PhysX card:


Marfia II GTX570 with 8800GS dedicated as PhysX processor:


Wow, I'm so surprised Looks like the 500 series benefits nicely from having a dedicated PhysX card just like the 400 series. Hmmm.....now where did I hear that before?



Oh and T_Flight, the 500 series cards don't have "wider RAM buses" than the cards they replace. GTX570 has a 320bit bus and GTX470 has a 320bit bus. GTX580 has a 384bit bus and GTX480 has a 384bit bus. And the only real thing they changed other than texture fill rate processing is they removed all the HPC parts from the GPU that were not needed for gaming which reduced transistor count from 3.2 billion to roughly 3 billion.

I can stay here and shoot holes in your ignorance all night, but honestly it's beginning to bore me

 
+1
I plan to run two 570's plus a dedicated physx card early this year. Its funny how you show hard facts but they don't provide any themselves except for plain numbers on the forum with no real screenshot...
 
Anyways, I have seen improvements what a physx card can do which is why I will be using one soon. Plus, it will add to my folding power.

Summary of BM in Metro 2033 w/Dedicated Physx

• Average Framerate: 30.00 SLI GTX 580 CPU Dedicated PhysX
• Average Framerate: 9.67 SLI GTX 580 GPU 1 Dedicated PhysX
• Average Framerate: 20.00 SLI GTX 580 GPU 2 Dedicated PhysX
• Average Framerate: 36.67 SLI GTX 580, GTX 285 (1gb ram) Dedicated PhysX
• Average Framerate: 34.33 SLI GTX 580, 8800GT (512 ram) Dedicated PhysX

 

• Average Framerate: 19.33 one GTX 580 CPU Dedicated PhysX
• Average Framerate: 20.00 one GTX 580 GPU1 Dedicated PhysX
• Average Framerate: 19.67 one GTX 580 GPU2 Dedicated PhysX
• Average Framerate: 19.67 one GTX 580 GTX 285 (1gb ram) Dedicated PhysX
• Average Framerate: 19.67 one GTX 580, 8800GT (512 ram) Dedicated PhysX

 
All BM’s done with my system specs in sig except I removed the Creative xFi PCIe card to allow for a Dedicated PhysX card.  I used an external Creative xFi USB 2 for all BM’s.
I chose to sum up with Average BM’s as I ran benches two times for each setting and the Average BM was ALWAYS the same while Min\Max varied greatly like 190 vs 155 on max and 6.08 vs 3.33 on min using same setup.
 
Fans ran @ 80% using MSI Afterburner, GPU's Vanilla speeds
Temp on any GTX 580 in any configuration was never above 62c
Temp on Dedicated GTX 285 never above 48c (rear exhaust fan)
Temp on Dedicated 8800GT never above 78c (middle mounted fan, not rear exhaust)
GTX 580 GPU 1 in PCIe Slot One
Dedicated PhysX GPU's in PCIe Slot Two
GTX 580 GPU 2 in PCIe Slot Three

 Two SLI GTX 580 CPU Dedicated PhysX
Driver 266.58266.58 ^WHQL
METRO 2033 BENCHMARK RESULTS
1/23/2011 1:30:55 PM
frontline

---

 
Options: Resolution: 2560 x 1600; DirectX: DirectX 11; Quality: Very High; Antialiasing: MSAA 4X; Texture filtering: AF 16X; Advanced PhysX: Enabled; Tesselation: Enabled; DOF: Disabled
Average Results

• Average Framerate: 30.00
• Max. Framerate: 215.05
• Min. Framerate: 6.50

 
Two SLI GTX 580 GPU1 Dedicated PhysX
Driver 266.58266.58 ^WHQL
 METRO 2033 BENCHMARK RESULTS
1/23/2011 12:54:38 PM
frontline

---

 
Options: Resolution: 2560 x 1600; DirectX: DirectX 11; Quality: Very High; Antialiasing: MSAA 4X; Texture filtering: AF 16X; Advanced PhysX: Enabled; Tesselation: Enabled; DOF: Disabled
Average Results

• Average Framerate: 9.67
• Max. Framerate: 158.96
• Min. Framerate: 1.05

 
Two SLI GTX 580 GPU2 Dedicated PhysX
Driver 266.58266.58 ^WHQL
 METRO 2033 BENCHMARK RESULTS
1/23/2011 1:02:07 PM
frontline

---

 
Options: Resolution: 2560 x 1600; DirectX: DirectX 11; Quality: Very High; Antialiasing: MSAA 4X; Texture filtering: AF 16X; Advanced PhysX: Enabled; Tesselation: Enabled; DOF: Disabled
 
Average Results

• Average Framerate: 20.00
• Max. Framerate: 188.18
• Min. Framerate: 6.01

 
Two SLI GTX 580, One GTX 285 (1gb ram) Dedicated PhysX
Driver 266.58266.58 ^WHQL
METRO 2033 BENCHMARK RESULTS
1/23/2011 11:01:59 AM
frontline

---

 
Options: Resolution: 2560 x 1600; DirectX: DirectX 11; Quality: Very High; Antialiasing: MSAA 4X; Texture filtering: AF 16X; Advanced PhysX: Enabled; Tesselation: Enabled; DOF: Disabled
AveGTrage Results

• Average Framerate: 36.67
• Max. Framerate: 160.98
• Min. Framerate: 4.13

 
Two SLI GTX 580, One 8800GT (512ram) Dedicated PhysX
Driver 266.58266.58 ^WHQL
METRO 2033 BENCHMARK RESULTS
1/23/2011 11:42:13 AM
frontline

---

 
Options: Resolution: 2560 x 1600; DirectX: DirectX 11; Quality: Very High; Antialiasing: MSAA 4X; Texture filtering: AF 16X; Advanced PhysX: Enabled; Tesselation: Enabled; DOF: Disabled
Average Results

• Average Framerate: 34.33
• Max. Framerate: 151.22
• Min. Framerate: 4.06

 
One GTX 580, CPU Dedicated PhysX
Driver 266.58266.58 ^WHQL
METRO 2033 BENCHMARK RESULTS
1/23/2011 12:28:07 PM
frontline

---

 
Options: Resolution: 2560 x 1600; DirectX: DirectX 11; Quality: Very High; Antialiasing: MSAA 4X; Texture filtering: AF 16X; Advanced PhysX: Enabled; Tesselation: Enabled; DOF: Disabled
Average Results
Average Results

• Average Framerate: 19.33
• Max. Framerate: 58.88
• Min. Framerate: 4.16

 
One GTX 580, GPU 1 GTX 580 Dedicated PhysX
Driver 266.58266.58 ^WHQL
METRO 2033 BENCHMARK RESULTS
1/23/2011 12:48:37 PM
frontline

---

 
Options: Resolution: 2560 x 1600; DirectX: DirectX 11; Quality: Very High; Antialiasing: MSAA 4X; Texture filtering: AF 16X; Advanced PhysX: Enabled; Tesselation: Enabled; DOF: Disabled
Average Results

• Average Framerate: 20.00
• Max. Framerate: 52.61
• Min. Framerate: 6.06

 
One GTX 580, One GPU 2 GTX 580 Dedicated PhysX
Driver 266.58266.58 ^WHQL
METRO 2033 BENCHMARK RESULTS
1/23/2011 11:19:23 AM
frontline

---

 
Options: Resolution: 2560 x 1600; DirectX: DirectX 11; Quality: Very High; Antialiasing: MSAA 4X; Texture filtering: AF 16X; Advanced PhysX: Enabled; Tesselation: Enabled; DOF: Disabled
Average Results
Average Results

• Average Framerate: 19.67
• Max. Framerate: 55.73
• Min. Framerate: 6.30

 
One GTX 580, One GTX 285 (1gb ram) Dedicated PhysX
Driver 266.58266.58 ^WHQL
METRO 2033 BENCHMARK RESULTS
1/23/2011 11:13:29 AM
frontline

---

 
Options: Resolution: 2560 x 1600; DirectX: DirectX 11; Quality: Very High; Antialiasing: MSAA 4X; Texture filtering: AF 16X; Advanced PhysX: Enabled; Tesselation: Enabled; DOF: Disabled
Average Results

• Average Framerate: 19.67
• Max. Framerate: 74.49
• Min. Framerate: 5.31

 
One GTX 580, One 8800GT(512 ram) Dedicated PhysX
Driver 266.58266.58 ^WHQL
METRO 2033 BENCHMARK RESULTS
1/23/2011 11:55:33 AM
frontline

---

 
Options: Resolution: 2560 x 1600; DirectX: DirectX 11; Quality: Very High; Antialiasing: MSAA 4X; Texture filtering: AF 16X; Advanced PhysX: Enabled; Tesselation: Enabled; DOF: Disabled
Average Results

• Average Framerate: 19.67
• Max. Framerate: 49.75
• Min. Framerate: 5.98

APEX Physics for clothing in Mafia II is irritating since it runs on the CPU.

So, I could test with dedicating one 570 to PhysX and then it won't revert to the CPU?

I'm sorry, I already ran it and I didn't take screenshots. I did take notes though.
 
All done with APEX High.
Sequence is Min FPS | Max FPS | Avg FPS
 
SLI and PhysX set to Auto
4.9 | 166.7 | 44.6
 
SLI and GPU2 as dedicated PhysX
5.2 | 142.9 | 52.4
 
SLI disabled and GPU2 as dedicated PhysX
5.3 | 166.7 | 52.9
 
I disabled SLI to see the results since it was mentioned here (linked from this document).
 
With every setup, when there's big explosions with lots of particles, all eight CPU cores max out.

For me, the goal was to optimize what I already have in my system. There is only two games that I play that have hardware accelerated PhysX, so I'm not going to get a card just for that.