Sign in with
Sign up | Sign in
Your question

PCI-E 1.x bandwidth saturation

Last response: in Graphics & Displays
Share
October 23, 2009 2:08:07 PM

Where can I get some reliable information on data transfer rates of graphics cards (i.e. bandwidth)?

I have pci-e 1.x (x16) slot and would like to buy a 2.x card, but don't want to buy more card than I have bandwidth, because I'm planning a major upgrade in 12-16 months.

I don't want to hear any questions on whether i've updated drivers and do i have an adequate power supply, etc., etc.. I simply want to know if the most recent cards are capable of saturating a pci-e 1.x (x16) slot (which for some reason I can't find any info on).
October 23, 2009 3:30:02 PM

@tomasz

that doesn't fully answer my question, but it is interesting to see that even today's biggest cards aren't hampered at half the speed. thanks.

anyone else?
m
0
l
Related resources
October 23, 2009 4:18:26 PM

just to clarify, quote from articles introduction:

It should also be noted that PCI-Express 2.0 doubles the bandwidth available per lane. So if you read about PCI-E x8 2.0 in this review, these results are representative of a PCI-E x16 link in 1.1 mode - like on many old motherboard.
m
0
l
October 23, 2009 4:47:08 PM

Quote:
So if you read about PCI-E x8 2.0 in this review, these results are representative of a PCI-E x16 link in 1.1 mode - like on many old motherboard.


Yes, I saw that. Still I'm hoping someone can point me in the direction of how much bandwidth newer cards need (i.e. how much raw data these cards are pulling through the interface). All I can find are articles from early 2008 saying that 2.x cards should be fine on 1.x slots, but yet nobody seems to have any numbers. It seems to me that newer cards (those since early 2008) may be starting to approach the capacity of 1.x, otherwise what's the point of 2.x and the upcoming 3.x and their greater transfer speeds?

I'm a numbers person because I like to do my own research. "It should be fine" doesn't do it for me.

Again, thank you tomasz for your efforts.
m
0
l
a c 130 U Graphics card
October 23, 2009 4:49:11 PM

The point isn't transfer speeds. It's how much power the lanes output.

1.0 - 75W
2.0 - 150W
3.0 - 300W?
m
0
l
October 23, 2009 5:48:48 PM

shadow187 said:
The point isn't transfer speeds. It's how much power the lanes output.

1.0 - 75W
2.0 - 150W
3.0 - 300W?


Wrong. A pci-e 1.x and 2.x slot only supply 75w. Additional power can be supplied with connectors up to 300w for 2.x.

Short of posting SIGs regs themselves, this article is pretty clear on the issue: http://www.10stripe.com/featured/quick/pci-express-2-0....

, and a simple google search on the subject will lead to other support. Yes, quicker transfer speeds allow for lower wattage = greater efficiency, but that is not the point.

The point is most definitely transfer speeds.
m
0
l
a c 130 U Graphics card
October 23, 2009 7:54:06 PM

Why should transfer speeds matter when we haven't even hit the ceiling of them? TSH showed 24 SSD drives, which had a capacity of 4GB/s transfer speeds. Yet the PCI-E 1.0x16 lane limited it to 2GB/s. If they'd used a PCI-E 2.0 x16 lane, then the ceiling wouldn't exist. But think about it; it takes 24 SSD drives to hit the ceiling of the PCI-E2.0 slot.

WHAT USER HAS 24 SSD DRIVES.

m
0
l
October 23, 2009 8:36:47 PM

shadow187 said:
Why should transfer speeds matter when we haven't even hit the ceiling of them? TSH showed 24 SSD drives, which had a capacity of 4GB/s transfer speeds. Yet the PCI-E 1.0x16 lane limited it to 2GB/s. If they'd used a PCI-E 2.0 x16 lane, then the ceiling wouldn't exist. But think about it; it takes 24 SSD drives to hit the ceiling of the PCI-E2.0 slot.

WHAT USER HAS 24 SSD DRIVES.


The article sounds interesting, what's the cite? Even still though, data transfer rates clearly have an effect - read the article tomasz cited above.

Also this:
http://www.tomshardware.com/reviews/pci-express-2.0,191...
m
0
l
a c 130 U Graphics card
October 23, 2009 8:54:47 PM

Luv, the Pci-E2.0x8 and PCI-E1.0x16 have a difference of 1 FPS in the benchmarks. In CoD4, the difference was 13, but they were both above 120FPS, so it mattered not.
m
0
l
!