PCI-SIC accelerates the creation of PCIe 5.0

[WYP]

PCI-SIC accelerates the creation of PCIe 5.0, with plans to complete the standard in 2019.



Read more on PCI-SIG's PCIe 5.0 standard.

 

[Piskeante]

totally unusefull since not even the 1080 ti can utilice PCIe 3.0 to max. This is just another tech announcement that serves for nothing.

 

[AngryGoldfish]

totally unusefull since not even the 1080 ti can utilice PCIe 3.0 to max. This is just another tech announcement that serves for nothing.

Emmm... graphics cards for games are not the only things that use PCI-e lanes.

 

[Piskeante]

Emmm... graphics cards for games are not the only things that use PCI-e lanes.

That's true, but if a GPU cannot max out PCIe 3.0, nor will do any of the other devices you may consider, even the fastest Nvme ssd's.

So, your statement those not change much of what i said.

 

[AngryGoldfish]

That's true, but if a GPU cannot max out PCIe 3.0, nor will do any of the other devices you may consider, even the fastest Nvme ssd's.

So, your statement those not change much of what i said.

The rate at which GPU's, SSD's, and other devices are growing is still quite fast. By 2019 there is no guarantee that add-in cards won't at least take somewhat advantage of the additional bandwidth and throughput. To claim that it is a waste makes little sense to me. The furthering of technology can often be seen as a waste initially, but that's because all technology doesn't advance at the same rate. That doesn't mean you should sit on your ass until someone else catches up. If the technology is there to be developed, develop it. The quicker they are to the gate the quicker it'll be initialised and integrated at a cost consumers and businesses can afford.

 

[xacid]

That's true, but if a GPU cannot max out PCIe 3.0, nor will do any of the other devices you may consider, even the fastest Nvme ssd's.

So, your statement those not change much of what i said.

People were arguing you stance when PCIe was first coming out. Cards of the day couldn't saturate an 8x AGP slot.

Giving any device that connects to a PCIe lane a lot of bandwidth id a good thing. Faster devices can talk to each other the better.

Dismissing this as unuseful is ignorant.

 

[Piskeante]

The rate at which GPU's, SSD's, and other devices are growing is still quite fast. By 2019 there is no guarantee that add-in cards won't at least take somewhat advantage of the additional bandwidth and throughput. To claim that it is a waste makes little sense to me. The furthering of technology can often be seen as a waste initially, but that's because all technology doesn't advance at the same rate. That doesn't mean you should sit on your ass until someone else catches up. If the technology is there to be developed, develop it. The quicker they are to the gate the quicker it'll be initialised and integrated at a cost consumers and businesses can afford.

i get your point. and i aggree. The thing is: PCIe 3.0 made it's very first appearance in 2009 with revision 0.5 It finally lauch in 2010 base 0.9. it's been more than 7 years since then, and not even the most powerfull GPU has ever been able to exceed or max out. In fact, many articles have been written about this issue, and the results are that a 1080ti in pcie 3.0 will only be like 1-2% improve from pcie 2.0

There is not a single review that has ever shown a noticeable increase in performance from 2.0 to 3.0. In fact, 3.0 is almost doubled the bandwith of 2.0 and still the differences is sometimes irrelevant.

Now, what i would like to point out is that, whenever (and this has been proved true for years) there is an advance in technology, even though is just not needed, once it's accepted by manufacturers, implies an increase in price.
So they are going to charge you more money to get a mobo pcie 5.0 which you´ll never use, not even in years now.

Moreover, consider that we are close to getting to silicon limits, so even if the process goes to 5nm, there will not be an advance in performance as we´ve seen lately unless we get a way to increase our performance just not by increasing core/mem speed.

People were arguing you stance when PCIe was first coming out. Cards of the day couldn't saturate an 8x AGP slot.

Giving any device that connects to a PCIe lane a lot of bandwidth id a good thing. Faster devices can talk to each other the better.

Dismissing this as unuseful is ignorant.

ignorant?? ok, so you are saying that the more speed a pcie has, the better speed talking to other devices. That's not true. The speed is not a problem for the already 3.0 but for the devices connected to the pcie 3.0 to achieve 15gb/s.

So a 200gb/s pcie 5x will increase the speed of the pheripherials connected to that bridge?? are you ing kiddin?¿??

who is the ignorant??? you. ok, that sounds better.

 

[Kaapstad]

ignorant?? ok, so you are saying that the more speed a pcie has, the better speed talking to other devices. That's not true. The speed is not a problem for the already 3.0 but for the devices connected to the pcie 3.0 to achieve 15gb/s.

So a 200gb/s pcie 5x will increase the speed of the pheripherials connected to that bridge?? are you ing kiddin?¿??

who is the ignorant??? you. ok, that sounds better.

I have managed to run PCI-E 3.0 out of bandwidth with graphics cards, it is not that difficult if you know how.

 

[Chopper3]

That's true, but if a GPU cannot max out PCIe 3.0, nor will do any of the other devices you may consider, even the fastest Nvme ssd's.

So, your statement those not change much of what i said.

Sorry but that's just not the case - I'm a server guy not a gaming guy and we have lots of things held back by PCie 3.0 speeds - specifically NVMe and Optane memory - the latter can easily flood PCIe 4.0 on pre-release servers using 'Purley+' Xeons.

 

[NeverBackDown]

This doesn't make sense. So they just made 4.0 useless and that releases next year? So they are saying in the span of one year, we are going from 32GB/s to 128GB/s?

 

[Avet]

PCIe 5.0 needs to exist and be available before upcoming technology can be developed for that standard. So yes it needs to precede everything that will be connected trough it, even though those parts don't exist yet.

 

[NeverBackDown]

PCIe 5.0 needs to exist and be available before upcoming technology can be developed for that standard. So yes it needs to precede everything that will be connected trough it, even though those parts don't exist yet.

That doesn't even make any sense

 

[Avet]

That doesn't even make any sense

It makes perfect sense. You can't create let's say SATA controloer, and develop drive around that if SATA interface isn't designed in the first place.

 

[NeverBackDown]

It makes perfect sense. You can't create let's say SATA controloer, and develop drive around that if SATA interface isn't designed in the first place.

You still don't make sense. What GPU is using PCI 5.0 interface? How is a GPU supposed to be made for this when it was just announced and the manufacturers are currently trying to make a 4.0 version that is due earlier than 5.0?

 

[Korreborg]

You need the know the exact specification before you can develop the hardware.
That the specification is done in 2019 does not mean that you will see MB with PCIe 5.0 in 2019.