PCI Express 3.0 x16 graphic card on PCI Express 2.0 x16 slot?

[softerist]

Hello,

I'm having a really difficult time about making a decision if I sould buy an GTX 780 3GB GDDR 5 384bit for an old motherboard (GA-Z68X-UD3-B3) with PCI Express 2.0 x16 slot.

I've read some articles on few forums but I'm not convinced if there will be a saturated performace because of the 2.0 slot.

If I look for the specs of the GTX 780 3GB GDDR 5 384bit compared to my old graphic card ASUS GTX 560 1GB 256bit I see this:

3,072 MB vs 1,024 MB
4,396 GFLOPS vs 1,243
288.4 GB/s vs 134.4 GB/s
45.8 GPixel/s vs 13 GPixel/s
166.4 GTexel/s vs 51.8 GTexel/s
6,008 MHz vs 4,200 MHz
2,304 vs 336
192 vs 56
38 dB vs 42.2 dB
250W vs 317W
1,502 MHz vs 1,050 MHz
384 bit vs 256 bit

But what does it mean the GTX 780 3GB GDDR 5 384bit on 2.0 slot?
Can someone give me the corect answer related to the limitations on this specs?
It is possible to lose some performace cause of a bottleneck?
Where can the bottleneck can apper for example?

It is possible a limtation for the transfer rate instead having 288.4 GB/s to get for example 200 GB/s or only 150?

What are the limitations of PCI Express 2.0 at x16?
What are the limitations of PCI Express 3.0 at x16?

Someone gave me that:

Base Clock Speed: PCIe 3.0 = 8.0GHz, PCIe 2.0 = 5.0GHz, PCIe 1.1 = 2.5GHz
Data Rate: PCIe 3.0 = 1000MB/s, PCIe 2.0 = 500MB/s, PCIe 1.1 = 250MB/s
Total Bandwidth: (x16 link): PCIe 3.0 = 32GB/s, PCIe 2.0 = 16GB/s, PCIe 1.1 = 8GB/s
Data Transfer Rate: PCIe 3.0 = 8.0GT/s, PCIe 2.0= 5.0GT/s, PCIe 1.1 = 2.5GT/s

but I can't understand it because there are other units and I can't associate them with the specs above to make an ideea.

Thank you!

 

[Hello man]

Hello,

I'm having a really difficult time about making a decision if I sould buy an GTX 780 3GB GDDR 5 384bit for an old motherboard (GA-Z68X-UD3-B3) with PCI Express 2.0 x16 slot.

I've read some articles on few forums but I'm not convinced if there will be a saturated performace because of the 2.0 slot.

If I look for the specs of the GTX 780 3GB GDDR 5 384bit compared to my old graphic card ASUS GTX 560 1GB 256bit I see this:

3,072 MB vs 1,024 MB
4,396 GFLOPS vs 1,243
288.4 GB/s vs 134.4 GB/s
45.8 GPixel/s vs 13 GPixel/s
166.4 GTexel/s vs 51.8 GTexel/s
6,008 MHz vs 4,200 MHz
2,304 vs 336
192 vs 56
38 dB vs 42.2 dB
250W vs 317W
1,502 MHz vs 1,050 MHz
384 bit vs 256 bit

But what does it mean the GTX 780 3GB GDDR 5 384bit on 2.0 slot?
Can someone give me the corect answer related to the limitations on this specs?

It is possible a limtation for the transfer rate instead having 288.4 GB/s to get for example 200 GB/s or only 150?

What are the limitations of PCI Express 2.0 at x16?
What are the limitations of PCI Express 3.0 at x16?

Someone gave me that:

Base Clock Speed: PCIe 3.0 = 8.0GHz, PCIe 2.0 = 5.0GHz, PCIe 1.1 = 2.5GHz
Data Rate: PCIe 3.0 = 1000MB/s, PCIe 2.0 = 500MB/s, PCIe 1.1 = 250MB/s
Total Bandwidth: (x16 link): PCIe 3.0 = 32GB/s, PCIe 2.0 = 16GB/s, PCIe 1.1 = 8GB/s
Data Transfer Rate: PCIe 3.0 = 8.0GT/s, PCIe 2.0= 5.0GT/s, PCIe 1.1 = 2.5GT/s

but I can't understand it because there are other units and I can't associate them with the specs above to make an ideea.

Thank you!

3.0 would be ideal, but PCI is backwards compatible, so there is no restriction, any card could work in any PCI Express slot. The performance decrease will not really be noticed, not with only one card. I would only say that you should not do it if the slot was not full bandwidth, like X8 or something like that.

 

[softerist]

Hello,

I'm having a really difficult time about making a decision if I sould buy an GTX 780 3GB GDDR 5 384bit for an old motherboard (GA-Z68X-UD3-B3) with PCI Express 2.0 x16 slot.

I've read some articles on few forums but I'm not convinced if there will be a saturated performace because of the 2.0 slot.

If I look for the specs of the GTX 780 3GB GDDR 5 384bit compared to my old graphic card ASUS GTX 560 1GB 256bit I see this:

3,072 MB vs 1,024 MB
4,396 GFLOPS vs 1,243
288.4 GB/s vs 134.4 GB/s
45.8 GPixel/s vs 13 GPixel/s
166.4 GTexel/s vs 51.8 GTexel/s
6,008 MHz vs 4,200 MHz
2,304 vs 336
192 vs 56
38 dB vs 42.2 dB
250W vs 317W
1,502 MHz vs 1,050 MHz
384 bit vs 256 bit

But what does it mean the GTX 780 3GB GDDR 5 384bit on 2.0 slot?
Can someone give me the corect answer related to the limitations on this specs?

It is possible a limtation for the transfer rate instead having 288.4 GB/s to get for example 200 GB/s or only 150?

What are the limitations of PCI Express 2.0 at x16?
What are the limitations of PCI Express 3.0 at x16?

Someone gave me that:

Base Clock Speed: PCIe 3.0 = 8.0GHz, PCIe 2.0 = 5.0GHz, PCIe 1.1 = 2.5GHz
Data Rate: PCIe 3.0 = 1000MB/s, PCIe 2.0 = 500MB/s, PCIe 1.1 = 250MB/s
Total Bandwidth: (x16 link): PCIe 3.0 = 32GB/s, PCIe 2.0 = 16GB/s, PCIe 1.1 = 8GB/s
Data Transfer Rate: PCIe 3.0 = 8.0GT/s, PCIe 2.0= 5.0GT/s, PCIe 1.1 = 2.5GT/s

but I can't understand it because there are other units and I can't associate them with the specs above to make an ideea.

Thank you!

3.0 would be ideal, but PCI is backwards compatible, so there is no restriction, any card could work in any PCI Express slot. The performance decrease will not really be noticed, not with only one card. I would only say that you should not do it if the slot was not full bandwidth, like X8 or something like that.

Thanky you for quick reply.

What do you mean by " it if the slot was not full bandwidth, like X8 or something like that"? Yes my MB have 2 PCIE 2.0 slots. One of x16 and one of x8, and ofc I will use the x16 slot.
I know that the PCIE slots are backwards compatible, the real question (for one card installed), is there a bottleneck? How bad is the bottleneck, does it worth (the money) or sould I just buy an 690 for example?

 

[17seconds]

You will have zero problems and no performance impact from using the 2.0 PCIe slot.

 

[Hello man]

Hello,

I'm having a really difficult time about making a decision if I sould buy an GTX 780 3GB GDDR 5 384bit for an old motherboard (GA-Z68X-UD3-B3) with PCI Express 2.0 x16 slot.

I've read some articles on few forums but I'm not convinced if there will be a saturated performace because of the 2.0 slot.

If I look for the specs of the GTX 780 3GB GDDR 5 384bit compared to my old graphic card ASUS GTX 560 1GB 256bit I see this:

3,072 MB vs 1,024 MB
4,396 GFLOPS vs 1,243
288.4 GB/s vs 134.4 GB/s
45.8 GPixel/s vs 13 GPixel/s
166.4 GTexel/s vs 51.8 GTexel/s
6,008 MHz vs 4,200 MHz
2,304 vs 336
192 vs 56
38 dB vs 42.2 dB
250W vs 317W
1,502 MHz vs 1,050 MHz
384 bit vs 256 bit

But what does it mean the GTX 780 3GB GDDR 5 384bit on 2.0 slot?
Can someone give me the corect answer related to the limitations on this specs?

It is possible a limtation for the transfer rate instead having 288.4 GB/s to get for example 200 GB/s or only 150?

What are the limitations of PCI Express 2.0 at x16?
What are the limitations of PCI Express 3.0 at x16?

Someone gave me that:

Base Clock Speed: PCIe 3.0 = 8.0GHz, PCIe 2.0 = 5.0GHz, PCIe 1.1 = 2.5GHz
Data Rate: PCIe 3.0 = 1000MB/s, PCIe 2.0 = 500MB/s, PCIe 1.1 = 250MB/s
Total Bandwidth: (x16 link): PCIe 3.0 = 32GB/s, PCIe 2.0 = 16GB/s, PCIe 1.1 = 8GB/s
Data Transfer Rate: PCIe 3.0 = 8.0GT/s, PCIe 2.0= 5.0GT/s, PCIe 1.1 = 2.5GT/s

but I can't understand it because there are other units and I can't associate them with the specs above to make an ideea.

Thank you!

3.0 would be ideal, but PCI is backwards compatible, so there is no restriction, any card could work in any PCI Express slot. The performance decrease will not really be noticed, not with only one card. I would only say that you should not do it if the slot was not full bandwidth, like X8 or something like that.

Thanky you for quick reply.

What do you mean by " it if the slot was not full bandwidth, like X8 or something like that"? Yes my MB have 2 PCIE 2.0 slots. One of x16 and one of x8, and ofc I will use the x16 slot.
I know that the PCIE slots are backwards compatible, the real question (for one card installed), is there a bottleneck? How bad is the bottleneck, does it worth (the money) or sould I just buy an 690 for example?

Sorry about not replying- The tracked thread panel has been broken for a while now, as well as the notifications pane. Just started working this morning. I meant that the x8 slot is slower, so a very fast card will be able to send less data across it. If it has to hold data back, it will use more on board GPU VRAM. Just use the x16 slot, and it will be fine.

 

[17seconds]

Not an issue at all.

http://www.pugetsystems.com/labs/articles/Impact-of-PCI-E-Speed-on-Gaming-Performance-518/

Conclusion
Our testing has pretty clearly shown that for gaming using either PCI-E 2.0 or PCI-E 3.0 will give you nearly identical performance. Oddly, in some benchmarks PCI-E 2.0 was actually faster than PCI-E 3.0. At the same time, x16 was not consistantly faster than x8. Again, x8 was actually faster than x16 in many cases.

http://www.hardocp.com/article/2010/08/16/sli_cfx_pcie_bandwidth_perf_x16x16_vs_x16x8/6#.U3QmPpUU-Uk

http://www.anandtech.com/show/5458/the-radeon-hd-7970-reprise-pcie-bandwidth-overclocking-and-msaa

 

[Roger Rogers-1382392]

Hello,
........................on 2.0 slot?
Can someone give me the corect answer related to the limitations on this specs?

What are the limitations of PCI Express 2.0 at x16?
What are the limitations of PCI Express 3.0 at x16?

but I can't understand it because there are other units and I can't associate them with the specs above to make an ideea.
Thank you!

......................... any card could work in any PCI Express slot. The performance decrease will not really be noticed, not with only one card. I would only say that you should not do it if the slot was not full bandwidth, like X8 or something like that.

I can see that in his case that is true, but it is not a panacea.

I have had a lot of trouble finding out that my (2006) XP computer will not allow me to run a HD7770 because that is beyond the `x16 Ver 2.0` spec; and my XP PC is x16 pre-Ver 2.0 spec.

I run an ATi Sapphire X1950 XTX 512Mb on it, and wanted a card with 1Gb at least; but I can only run something like a HD4770 (Version 2.0 x16). However, I had bought two HD7770`s by the time I found that out and neither will work; no matter which drivers are installed.

600W PSU.

 

[AIRCRAFTWOODY]

This might shed a little light
http://www.tested.com/tech/457440-theoretical-vs-actual-bandwidth-pci-express-and-thunderbolt/

 

[voodue]

PCI express 2.0 at x16 compared to 3.0 x16 is noticeable. I spent two weeks deciding on my latest setup, 1: upgrade and system with pcie2.0 or go for the has well shhafordable setup, once I got the GPU installed into my new haswell board I couldn't believe there was data supporting 2.0 x16 I later got second GPU setting the PCI e ports to x8 each and the difference between x8 sli 760s and x16 was measurable both3.0 standard on two different motherboards, will post more info asap not sure I have screen caps anymore.


To be clear, i went from AMD 2.0 x16 single gtx 760 to 3.0x16 and saw a jump in frame rate and a noticeable difference in smoothness. After going from x16 3.0 to dual gpus in x8 3.0 i didnt notice a huge jump. I later bought a z87-ws asus board and again with both gpu's in x16 3.0 ports i saw more frames per second..

The games i used were BF4
Call of duty ghosts
League of legends

Unfortunately all i have of the old setups are some pictures

(X16 3.0 dual gpu's Liquid cooled http://imgur.com/LPS21EH http://imgur.com/akNK5qJ


Air cooled http://imgur.com/JnYDl8S http://imgur.com/y0WMuWj)

Single card at x16 3.0 ( when i noticed a frame rate difference but apparently did not record it :/)
http://imgur.com/N8D5Dll


And finally the rats nest Amd quad core 2.0 x16
http://imgur.com/wkcjDP8

 

[17seconds]

Here's a recent article.
http://www.techpowerup.com/reviews/NVIDIA/GTX_980_PCI-Express_Scaling/

 

[Duncan Tallon]

Not an issue at all.

http://www.pugetsystems.com/labs/articles/Impact-of-PCI-E-Speed-on-Gaming-Performance-518/

Conclusion
Our testing has pretty clearly shown that for gaming using either PCI-E 2.0 or PCI-E 3.0 will give you nearly identical performance. Oddly, in some benchmarks PCI-E 2.0 was actually faster than PCI-E 3.0. At the same time, x16 was not consistantly faster than x8. Again, x8 was actually faster than x16 in many cases.

http://www.hardocp.com/article/2010/08/16/sli_cfx_pcie_bandwidth_perf_x16x16_vs_x16x8/6#.U3QmPpUU-Uk

http://www.anandtech.com/show/5458/the-radeon-hd-7970-reprise-pcie-bandwidth-overclocking-and-msaa

Wow, this is really interesting.

 

[17seconds]

Another recent review of PCIe lane scaling:
http://www.guru3d.com/articles-pages/pci-express-scaling-game-performance-analysis-review,1.html