-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
Intel 10th Gen Comet Lake thread
- Thread starter AthlonXP1800
- Start date
Soldato
Can we get some review much newer than from a year ago? Things have changed. AGESA 1004 & 1005 have tweaked pcie 4.0.
Ryzen 3000 CPUs have received proper boosting, let alone the 12+ drivers we got since.
Also we do not know how WDDM 2.7 would affect this, given it has a lot of AMD custom enhancements. (since WDDM 2.7 will be used on next Xbox)
Ofc Nvidia would also have Ampere at pcie 4.0, any attempt to stick to 3.0 would look stupid.
That would make perfect sense as anything to take the shine off what is a dubious product (even if it does take gaming performance crown) would be a good idea. AMD does still offer shockingly better value and much lower power consumption as well as more than acceptable gaming performance.
Associate
- Joined
- 28 Sep 2018
- Posts
- 2,259
You would need a significant jump in performance to come close to pcie3.0 x16 saturation.
https://www.techpowerup.com/review/nvidia-geforce-rtx-2080-ti-pci-express-scaling/
you can see here that a doubling of PCIe bandwidth nets a 2-3% gain on average. Again that’s with doubling of the bandwidth.
DP 2.0 and hdmi 2.1 is much more relevant to a consumer than the bus interface.
Did you not bother actually reading the conclusions?
"It also produces strong evidence that PCIe 4.0 won't be needed for even more powerful next-gen graphics cards because our three tested resolutions reveal more details."
"If you look closely, you'll notice that lower resolutions show bigger differences in performance when changing the PCI-Express bandwidth, which seems counter-intuitive at first. Doesn't the graphics card work harder at higher resolutions? While that may (mostly) be true, graphics card load does not increase PCI-Express bandwidth; it actually lowers it because frame rates are lower. The amount of data transferred over the PCIe bus is fairly constant—per frame. So if the graphics card can run at higher FPS rates because the resolution is lower, the PCIe bus does have more traffic moving across it."
And why would drivers have any significant bearing on PCIe utilisation? Where is the data to support this?
All available tests and evidence clearly show that PCIe bandwidth matters very little, even to the highest performing cards and that there is no reason at all to believe this would change with the next generation of cards.
As the article above clearly explains, the PCIe bandwidth only affects the "instructions" flowing from the CPU to the GPU which, at a given framerate, is fairly constant regardless of the GPU.
Of course because it's good marketing and yes, regardless of whether it actually provides any real benefit, not supporting the latest standards does look a bit odd.
The problem with PCIe 4 is that it's 90% marketing. It has no real-world benefit outside raw sequential throughput of SSDs, which matters very little to the vast majority of people.
Soldato
Bandwidth matters more in some rendering/professional programs its not all about Gaming. PCIE 4.0 matters much more in proper compute environment.
Think 2080TI can just about max out 8X PCIE 3.0 so even 3080TI unlikely to max out 16X in gaming but again tied to frame rate if low res and very high fps then its worst case scenario, unlikely in real world to be still on 1080P with such a card.
Will be interesting to see though if any difference, hopefully they can do 1% and 0.1% this time
Think 2080TI can just about max out 8X PCIE 3.0 so even 3080TI unlikely to max out 16X in gaming but again tied to frame rate if low res and very high fps then its worst case scenario, unlikely in real world to be still on 1080P with such a card.
Will be interesting to see though if any difference, hopefully they can do 1% and 0.1% this time
Associate
- Joined
- 13 Mar 2009
- Posts
- 704
Well the chip I have at the moment is 60 quid more than when I bought it, I take it these strange weird and wonderful price dips and highs are just down to demand/supply?
I think you should be asking if PCIE4 is as pointless for GPUs as you believe, why has Intel included a tool on z490 boards to overclock the PCIE3 bus.
Why overclock pcie3 bus if next gen GPUs dont need it
I'd hesitate before placing too much faith in one older Techpowerup article testing a mid range product.
That 1-3% will continue to scale slowly as more powerful GPU's with Pcie 4.0 get released.
I'm not saying that Pcie 3.0 isn't enough. It clearly is enough and will be for a good while yet...
But people worry about future proofing I suppose.
That 1-3% will continue to scale slowly as more powerful GPU's with Pcie 4.0 get released.
I'm not saying that Pcie 3.0 isn't enough. It clearly is enough and will be for a good while yet...
But people worry about future proofing I suppose.
PCI-E isn't just graphics, NVMEs are big in enterprise solutions now, especially for media work. There's also the Intel Octane (is that what it's called?) that runs on PCI-E. PCI-E 4 won't be needed for graphics, but there's a number of other things that could use the extra bandwidth.
Last edited:
It's tangentially related! CPUs have to support new hardware standards. So PCI-E 4 is only going to happen if Intel include hardware support on their CPU's and it's supported by the chipset.
The discussion is just covering whether or not Intel including PCI-E 4 support is useful or not.
Associate
- Joined
- 28 Sep 2018
- Posts
- 2,259
PCIe 4.0 is absolutely useful and the to advance technology forward you need to move the whole eco system along.
As you and others mentioned, for storage and peripherals in the professional space it can have a number of key benefits. However thinking a 3080ti is going to be bottlenecked without it is the wrong hill to fight on.
Looking forward to the 20th and seeing what we get from OC results for CPU and mem.
https://www.google.co.uk/amp/s/www....t-at-least-you-can-insure-it-for-overclocking
Overclocking insurance
Overclocking insurance
Soldato
- Joined
- 22 Nov 2018
- Posts
- 2,715
Some people upgrade their GPU 2 or 3 times though. So somebody with a 3900x might get a 3080ti this year then a 4080ti in 2022 followed by a 5080ti in 2024. Surely a 5080ti will saturate PCI-E 4.
Today's high end CPUs will surely last at least 4 years, even though stagnation is over.
Everybody's talking about the 3080ti like it's the last GPU that nvidia will make.
Today's high end CPUs will surely last at least 4 years, even though stagnation is over.
Everybody's talking about the 3080ti like it's the last GPU that nvidia will make.
this is the cpu forum right.
Intel will need to crash the prices, such a bad time for them. Architectural security flaws, heat, a cheaper competition currently trading blows and expected to do better at the end of the year plus all in the middle of a pandemic with high demand for hardware.
Intel will need to crash the prices, such a bad time for them. Architectural security flaws, heat, a cheaper competition currently trading blows and expected to do better at the end of the year plus all in the middle of a pandemic with high demand for hardware.
Soldato
- Joined
- 22 Nov 2018
- Posts
- 2,715
We've been here before though. Pentium 4 started off hot and slower than the competition. They didn't need to cut prices though because people bought them anyway because of the brand.
Intel are always faster so the benchmarks must be inaccurate etc.
Soldato
PCI-E 4.0 won't be a thing for me until at least 2022. Could have had PCI-E 4.0,but AMD said B450 was not suitable,so I will have to stick PCI-E 3.0,and a Zen2 at best! 
Intel owners have to buy a new motherboard every year, so perhaps it's no surprise they don't care what happens in future years