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

GTX 680 for 4,1 recommendations

degl

degl

macrumors member
Original poster
Jun 12, 2010
49
1
Caracas
Hi folks;

I'm looking to replace my 1.3gb GTX 570 with a 4GB GTX 680, but i don't know if this feasible, what about the molex connector? been 6 pin the used on Mac Pro and for what i can see the GTX 680 came with a 8 pin molex connector. What model do you guys recommend?

Thanks in advance.
 
Several GTX680s come with Dual 6 pins.

Even the EVGA FTW 680 4GB has a 6 & 8, but runs with just dual 6s.

We have EFI versions available or you can run unflashed ones.
 
Hi MacVidCard, you mean that I could install the EVGA GeForce GTX 680 + 4GB w / Backplate in my mac pro? even if it has a 6 pin connector and one 8-pin connector? For intensive use with Octane render.
 
degl said:
And what type of cables can be those?
Thanks
Click to expand...

6 pin cables, i.e the ones you need with any video card. The 6 pin fits into the 8 pin port on the video card, leaving two pins open.

Works fine for me. I played through Sleeping Dogs on my bootcamp partition without any problems, and windows is less power efficient on MP than OS X.

I have had zero reliability issues (so far). I have the luxury of doing no mission critical stuff on my Mac. And I have multiple backups solutions.
 
My 4GB MSI GTX 680 card has two 6-pin connections.
IMHO, if you are only going to be using it with OS X, then the 4GB MSI 680 is fine.
If you are going to use it with Windows under Bootcamp, then get a 2GB 680 not a 4GB.
 
Zotac has a 4GB 680 with dual 6-pin connectors. Works like a champ in my 5,1.
 
Go with 4GB if you'r using Octane....

Topper said:
My 4GB MSI GTX 680 card has two 6-pin connections.
IMHO, if you are only going to be using it with OS X, then the 4GB MSI 680 is fine.
If you are going to use it with Windows under Bootcamp, then get a 2GB 680 not a 4GB.
Click to expand...


But using Octane this would cut the possible scene size in half. Only half as many polys could be loaded.

I say definitely go with 4GB if you'r using Octane....

When I got my first EVGA FTW+ 680 4GB it had a 6 & 8 pin connector and I ran it fine for a while on two 6 pin cables. When I got my second card I of course installed a 450w aux psu.
 
The FTW 680 4GB is designed to work with only 2 @ 6 pins if that is what you have.

It offers the ability to include 8 pin if you have it but not required.
 
MacVidCards said:
The FTW 680 4GB is designed to work with only 2 @ 6 pins if that is what you have.

It offers the ability to include 8 pin if you have it but not required.
Click to expand...

Interesting. I have an EVGA GTX 680+ 4 GB that I have been using with 2x 6 pin connectors (it is supposed to have a 6 + 8). Maybe my card has the same requirements as the FTW 680.

I work my card _HARD_ when wife lets me, I'll play a game for 10 hours straight on a cold dreary Sunday. Got to get through that steam backlog!
 
Topper said:
My 4GB MSI GTX 680 card has two 6-pin connections.
IMHO, if you are only going to be using it with OS X, then the 4GB MSI 680 is fine.
If you are going to use it with Windows under Bootcamp, then get a 2GB 680 not a 4GB.
Click to expand...

Why would you not get the 4GB if you're going to use it in Windows?
 
grue said:
Why would you not get the 4GB if you're going to use it in Windows?
Click to expand...

A 4GB 680 only runs at PCIe 1.1 bus speeds under Windows(Bootcamp) even if you flash it.
A "flashed" 2GB 680 operates at PCIe 2.0 bus speeds under Windows(Bootcamp).
 
Topper said:
A 4GB 680 only runs at PCIe 1.1 bus speeds under Windows(Bootcamp) even if you flash it.
A "flashed" 2GB 680 operates at PCIe 2.0 bus speeds under Windows(Bootcamp).
Click to expand...

Oooooooh. That's a pretty pro thing to know. I haven't had a lot of luck with my search-fu (Netkas' page is down), how involved is the flashing procedure, do you have a link?

edit: Nevermind, got the cached version.

For anyone else who needs it:

http://webcache.googleusercontent.com/search?q=cache:http://forum.netkas.org/index.php?topic=5709.0
 
Topper said:
A 4GB 680 only runs at PCIe 1.1 bus speeds under Windows(Bootcamp) even if you flash it.
A "flashed" 2GB 680 operates at PCIe 2.0 bus speeds under Windows(Bootcamp).
Click to expand...

It's time we put this misunderstanding to rest.

When flashed with a rom written for it, a 680 (or any Nvidia 6xx card, for that matter) WILL run at PCIE 2.0 in both OS's.

I think part of the misunderstanding has to do with the way this functions now.

In both OSX and Windows, the card runs at PCIE 1.0 until needed, at which point it switches modes into 2.0. So the only way to be 100% certain is to use a test that places a load on the card.

In Windows via Bootcamp, GPU-Z will show a card as being capable of 2.0 but running at 1 UNTIL YOU LOAD THE RENDER TEST.

Have a look at the images, these are from our GTX680 4GB running in a Mac Pro in Windows via Bootcamp. Note that GPU-Z only shows that card running at 1.0 to begin with, then running the render test it kicks in to 2.0. The render test is gotten to by hitting the "?" just to the right of the PCIE field.

You can observe the same behavior in OSX using "lspci" (Google is your friend)

Note that "LnkCap" shows 5.0 in both, but in first iteration, it shows "LnkSta" at 2.5, I then ran CUDA-Z which kicked the speed up for the second time where you see LnkSta at 5.0


01:00.0 VGA compatible controller: nVidia Corporation Unknown device 1180 (rev a1) (prog-if 00 [VGA controller])
Subsystem: nVidia Corporation Unknown device 102e
Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+
Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
Latency: 0
Interrupt: pin A routed to IRQ 18
Region 0: Memory at 9e000000 (32-bit, non-prefetchable)
Region 1: Memory at <ignored> (64-bit, prefetchable)
Region 3: Memory at <ignored> (64-bit, prefetchable)
Region 5: I/O ports at 3000 [disabled]
Capabilities: [60] Power Management version 3
Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
Status: D0 PME-Enable- DSel=0 DScale=0 PME-
Capabilities: [68] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable+
Address: 00000000fee00000 Data: 4092
Capabilities: [78] Express (v2) Endpoint, MSI 00
DevCap: MaxPayload 256 bytes, PhantFunc 0, Latency L0s unlimited, L1 <64us
ExtTag+ AttnBtn- AttnInd- PwrInd- RBE+ FLReset-
DevCtl: Report errors: Correctable- Non-Fatal- Fatal- Unsupported-
RlxdOrd+ ExtTag+ PhantFunc- AuxPwr- NoSnoop+
MaxPayload 128 bytes, MaxReadReq 512 bytes
DevSta: CorrErr- UncorrErr- FatalErr- UnsuppReq- AuxPwr- TransPend-
LnkCap: Port #4, Speed 5GT/s, Width x16, ASPM L0s L1, Latency L0 <512ns, L1 <4us
ClockPM+ Suprise- LLActRep- BwNot-
LnkCtl: ASPM Disabled; RCB 64 bytes Disabled- Retrain- CommClk-
ExtSynch- ClockPM+ AutWidDis- BWInt- AutBWInt-
LnkSta: Speed 2.5GT/s, Width x16, TrErr- Train- SlotClk+ DLActive- BWMgmt- ABWMgmt-
Capabilities: [b4] Vendor Specific Information <?>
Capabilities: [100] #10de
Capabilities: [118] #3f

01:00.0 VGA compatible controller: nVidia Corporation Unknown device 1180 (rev a1) (prog-if 00 [VGA controller])
Subsystem: nVidia Corporation Unknown device 102e
Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+
Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
Latency: 0
Interrupt: pin A routed to IRQ 18
Region 0: Memory at 9e000000 (32-bit, non-prefetchable)
Region 1: Memory at <ignored> (64-bit, prefetchable)
Region 3: Memory at <ignored> (64-bit, prefetchable)
Region 5: I/O ports at 3000 [disabled]
Capabilities: [60] Power Management version 3
Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
Status: D0 PME-Enable- DSel=0 DScale=0 PME-
Capabilities: [68] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable+
Address: 00000000fee00000 Data: 4092
Capabilities: [78] Express (v2) Endpoint, MSI 00
DevCap: MaxPayload 256 bytes, PhantFunc 0, Latency L0s unlimited, L1 <64us
ExtTag+ AttnBtn- AttnInd- PwrInd- RBE+ FLReset-
DevCtl: Report errors: Correctable- Non-Fatal- Fatal- Unsupported-
RlxdOrd+ ExtTag+ PhantFunc- AuxPwr- NoSnoop+
MaxPayload 128 bytes, MaxReadReq 512 bytes
DevSta: CorrErr- UncorrErr- FatalErr- UnsuppReq- AuxPwr- TransPend-
LnkCap: Port #4, Speed 5GT/s, Width x16, ASPM L0s L1, Latency L0 <512ns, L1 <4us
ClockPM+ Suprise- LLActRep- BwNot-
LnkCtl: ASPM Disabled; RCB 64 bytes Disabled- Retrain- CommClk-
ExtSynch- ClockPM+ AutWidDis- BWInt- AutBWInt-
LnkSta: Speed 5GT/s, Width x16, TrErr- Train- SlotClk+ DLActive- BWMgmt- ABWMgmt-
Capabilities: [b4] Vendor Specific Information <?>
Capabilities: [100] #10de
Capabilities: [118] #3f


I can not actually speak for any other 680 4GB cards but ours run at 2.0 in both OS's. They do not always show this from System profiler, but do if you know where to check. Also, any of the CUDA-Z type benchmarks or OpenCl OceanWaves bandwidth will also verify the actual speed.
 

Attachments

  • 1.0.PNG
    1.0.PNG
    36.1 KB · Views: 141
  • 1.1.PNG
    1.1.PNG
    52.6 KB · Views: 121
  • 2.0.PNG
    2.0.PNG
    301.3 KB · Views: 133
Last edited:
MacVidCards said:
When flashed with a rom written for it, a 680 (or any Nvidia 6xx card, for that matter) WILL run at PCIE 2.0 in both OS's.
Click to expand...

Interesting, thank you.
Guess I've got to flash my card.

Oops, can't find rom for my MSI 4GB GTX 680. Looks like I'm stuck on PCIe 1.1.
 
Last edited:
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back