New PCI-E standards
New PCI-E standards
https://www.tomshardware.com/news/what- ... 39063.html
https://www.techspot.com/news/78355-pci ... aunch.html
Wonder how long for NVIDIA to catch up on PCI-E 4.0 (New AMD Navi cards are supposed to have it)
https://www.techspot.com/news/78355-pci ... aunch.html
Wonder how long for NVIDIA to catch up on PCI-E 4.0 (New AMD Navi cards are supposed to have it)
Re: New PCI-E standards
Even if we have PCI-E 4.0 Ready MB, with PCI-E 3.0 devices there are no actual improvement in bandwidth. I'm planning to buy Ryzen 3900x and considering between x470 and x570 chipsets. Probably I should wait for comparisons, x470 most likely will perform worse.
PC: RTX 2070 | Ryzen R9 5950X (no OC) | 64 GB RAM
Notebook: RTX 4060 | Ryzen R9 7945HX | 32 GB RAM
Notebook: RTX 4060 | Ryzen R9 7945HX | 32 GB RAM
Re: New PCI-E standards
Ryzen 9 series supports PCI-E 4
x570 chipset also supports it
The SSDs and such will follow soon after the release of these two next month
Some "preliminary" ones announced already
NVIDIA will catch up (I am sure) to maintain the advantage
https://www.tomshardware.com/news/ssd-p ... 38418.html
https://www.tomshardware.com/news/patri ... 39482.html
https://www.engadget.com/2019/05/31/gig ... pcie4-ssd/
Why not the 3950X
x570 chipset also supports it
The SSDs and such will follow soon after the release of these two next month
Some "preliminary" ones announced already
NVIDIA will catch up (I am sure) to maintain the advantage
https://www.tomshardware.com/news/ssd-p ... 38418.html
https://www.tomshardware.com/news/patri ... 39482.html
https://www.engadget.com/2019/05/31/gig ... pcie4-ssd/
Why not the 3950X
Re: New PCI-E standards
I'm also in the (lengthy) process of acquiring bits and upgrading to 3900X / ASUS Crosshair VIII Hero X570. Can't find a board nor the cpu in Aus for love nor money at the moment. The 3900X seemed to be in a sweet spot for Mhz vs cores for non-overclockers.
I spotted a review of a PCIE4 nvme ssd last week (lost the link), the result was only a marginal increase in real-world performance.
I spotted a review of a PCIE4 nvme ssd last week (lost the link), the result was only a marginal increase in real-world performance.
I really do like it here.
Re: New PCI-E standards
Depends on "your real world" usage.
Still early in development at this time
https://www.engadget.com/2019/05/31/gig ... pcie4-ssd/
https://www.tomshardware.com/reviews/ph ... ,6173.html
Still early in development at this time
https://www.engadget.com/2019/05/31/gig ... pcie4-ssd/
https://www.tomshardware.com/reviews/ph ... ,6173.html
Re: New PCI-E standards
and
https://www.techpowerup.com/review/pci- ... 000/4.html
https://www.techpowerup.com/review/pci- ... 000/4.html
We were really curious to see what performance gains can be enjoyed when using a blazing-fast PCI-Express 4.0 NVMe SSD, so we ran extensive tests using actual real-life applications, not just synthetic benchmarks that don't take into account how applications work today.
While the synthetic numbers are truly impressive for PCI-Express 4.0, showing over 5 GB/s actual transfer rates, which is 43% faster than the same SSD running at PCI-Express 3.0, the actual real-life gains are small.
Averaged across our test suite, we only see a 1% improvement, nothing you'd ever notice in real-life.
I really do like it here.
Re: New PCI-E standards
I'm interested in it not for SSD stuff but for improved bandwidth CPU<->GPU.
Re: New PCI-E standards
Presently only AMD/ATI cards support the new 4.0 version
Guess we have to wait for the next version of NVidia cards.
I believe that theoretically it should double the bandwidth
Guess we have to wait for the next version of NVidia cards.
I believe that theoretically it should double the bandwidth
Re: New PCI-E standards
I was going to avoid replying and just let it die, but I read the page you linked to and I have to say... you gotta be kidding.hydra3333 wrote: ↑Tue Jul 30, 2019 5:10 amand
https://www.techpowerup.com/review/pci- ... 000/4.htmlWe were really curious to see what performance gains can be enjoyed when using a blazing-fast PCI-Express 4.0 NVMe SSD, so we ran extensive tests using actual real-life applications, not just synthetic benchmarks that don't take into account how applications work today.
While the synthetic numbers are truly impressive for PCI-Express 4.0, showing over 5 GB/s actual transfer rates, which is 43% faster than the same SSD running at PCI-Express 3.0, the actual real-life gains are small.
Averaged across our test suite, we only see a 1% improvement, nothing you'd ever notice in real-life.
Installing Itunes is part of the test "real life suite"
I am sorry but how many times does anybody installs that. Personally its zero for me
The ADATA is given as 1TB but the other drive is a 2TB size so are we comparing scan times on what.
Like I said its still early but reviews like that are iffy, installing ITunes is a real life app? Because we do that twice a day.
Re: New PCI-E standards
Zero for me too.
Re: New PCI-E standards
Well, er, now that you point it out ... I was looking more at their purported result rather than their test regime Lazy me, having had a background in testing stuff. Oh well.
I really do like it here.
Re: New PCI-E standards
The websites themselves might be reliable, but some of the reviewers might be freelancers.
They will review anything to earn money. They might mean well but...
That reviewer might be an avid Itunes user and to him it is real life usage, just not to many others.
Installing software requires cpu usage, storage read/write, ram, and A/V interaction
If your bench-marking an SSD you should just move data, of various sizes, from 1 to 2 or 1 to 1 to see speeds.
OS loading tests depend on how many apps are installed and how many are allowed to load at start up, again not a good test.
Also, If you save 0.5ms on a 5ms task it is not impressive, but if you save 10sec on a 100sec task that isn't bad.
They are both 10%
They will review anything to earn money. They might mean well but...
That reviewer might be an avid Itunes user and to him it is real life usage, just not to many others.
Installing software requires cpu usage, storage read/write, ram, and A/V interaction
If your bench-marking an SSD you should just move data, of various sizes, from 1 to 2 or 1 to 1 to see speeds.
OS loading tests depend on how many apps are installed and how many are allowed to load at start up, again not a good test.
Also, If you save 0.5ms on a 5ms task it is not impressive, but if you save 10sec on a 100sec task that isn't bad.
They are both 10%