PCI-SIG Clarifies: No 300W Power Boost

[patrickjp93]

http://www.guru3d.com/news-story/pci-express-4-will-deliver-75-watt-over-pcie-not-300w.html

 

The short of it? Tom's Hardware (and many others) got their analysis wrong. It's 75W per slot, same as always, and 75*4 = 300, just like I said when the spec launched.

 

Quote

Update, 8/24/16, 2:06pm PT:PCI-SIG reached out to tell us that the power increase for PCI Express 4.0 will come from secondary connectors and not from the slot directly. They confirmed that we were initially told incorrect information. We have redacted a short passage from our original article that stated what were originally told, which is that the slot would provide at least 300W, and added clarification:

PCIe 3.0 max power capabilities: 75W from CEM + 225W from supplemental power connectors = 300W total

PCIe 4.0 max power capabilities: TBD

New value “P” = 75W from CEM + (P-75)W from supplemental power connectors.

 

Edited August 25, 2016 by Godlygamer23
Clean up.

[DXMember]

5 minutes ago, patrickjp93 said:

The short of it? Tom's Hardware (and many others) got their analysis wrong.

Quote

They confirmed that we were initially told incorrect information.

I don't think either of them did analysis they just reported the marketing material presented to them

[patrickjp93]

2 minutes ago, DXMember said:

I don't think either of them did analysis they just reported the marketing material presented to them

It's due diligence to actually read the spec. When did our standards for tech journalism fall this far?

[DXMember]

3 minutes ago, patrickjp93 said:

It's due diligence to actually read the spec. When did our standards for tech journalism fall this far?

well believing it would do 300W would be nuts... I have to agree on that

if it were true it would have to completely rewire the PCI-E slot which then wouldn't be backwards compatible

because no way in hell would they pull 25A on those tiny 5 pins, lol

 

but I still don't understand the 300W... are they referring to the external power plugs? The PCI-E 8-pin? What prevents them from making another 290x Lightning with 3x 8-pins?? That'd be more than 300W or is that outside spec?

[patrickjp93]

2 minutes ago, DXMember said:

well believing it would do 300W would be nuts... I have to agree on that

if it were true it would have to completely rewire the PCI-E slot which then wouldn't be backwards compatible

because no way in hell would they pull 25A on those tiny 5 pins, lol

 

but I still don't understand the 300W... are they referring to the external power plugs? The PCI-E 8-pin? What prevents them from making another Lightning with 3x 8-pins?? That'd be more than 300W or is that outside spec?

It's 300W of power over the entire bus at a maximum of 75W per slot for 4 slots.

[DXMember]

Just now, patrickjp93 said:

It's 300W of power over the entire bus at a maximum of 75W per slot for 4 slots.

oh, I get it now

 

But they would still have to use supplementary connectors, no? Do they intend to get all that from 24-pin?

[Doobeedoo]

That looks more reasonable. 

[Curufinwe_wins]

Lol, a more aggressive PJP has appeared!

[patrickjp93]

48 minutes ago, DXMember said:

oh, I get it now

 

But they would still have to use supplementary connectors, no? Do they intend to get all that from 24-pin?

Not for the base 300W.

 

22 minutes ago, Curufinwe_wins said:

Lol, a more aggressive PJP has appeared!

It's my Aussie side

[juretrn]

2 hours ago, patrickjp93 said:

It's due diligence to actually read the spec. When did our standards for tech journalism fall this far?

At least in the mainstream media, there never were any. 

[Lays]

2 hours ago, patrickjp93 said:

It's due diligence to actually read the spec. When did our standards for tech journalism fall this far?

It's probably because of all the tech forums that give all those shitty tech journalist sites & rumors so much publicity.

 

They're mainly just taking advantage of it, there's always boatloads of rumors floating around the tech world, and as long as people keep discussing these insanely unrealistic rumors, people will most likely continue to have crappy tech "news" sites that take advantage of the discussions.

[BuckGup]

Wow wrong again. This has been really happening a lot from Tom's Hardware

[Rune]

Aw...I wanted a clean looking dual chip card monstrosity...

[Yog]

2 hours ago, patrickjp93 said:

It's due diligence to actually read the spec. When did our standards for tech journalism fall this far?

I imagine at the same time every other type of journalism did as well.

[PopReference]

They really shouldn't have said that in their article but I don't blame them for the misunderstanding. The exact specs for PCIe are not easy to find, at least I couldn't find from searching, and what some basic spec to be found does say is "300 watts" even for 2.0(if I remember right) with no really distinguishing between total or per connection.

 

I think this falls more into the PCIe standard just not sharing enough information + some lazy journalism.

[patrickjp93]

20 minutes ago, PopReference said:

They really shouldn't have said that in their article but I don't blame them for the misunderstanding. The exact specs for PCIe are not easy to find, at least I couldn't find from searching, and what some basic spec to be found does say is "300 watts" even for 2.0(if I remember right) with no really distinguishing between total or per connection.

 

I think this falls more into the PCIe standard just not sharing enough information + some lazy journalism.

Search PCI SIG technical specifications. Or, go to their website and do some searching. It's publicly available.

[PopReference]

27 minutes ago, patrickjp93 said:

Search PCI SIG technical specifications. Or, go to their website and do some searching. It's publicly available.

I looked now and i looked then and the first thing I find is a login screen. How about you do the searching and tell me about it.

 

The original specs are hard to find for most people is my point.

[patrickjp93]

3 minutes ago, PopReference said:

I looked now and i looked then and the first thing I find is a login screen. How about you do the searching and tell me about it.

 

The original specs are hard to find for most people is my point.

12 seconds of searching.

https://pcisig.com/specifications?field_technology_value[]=express&field_revision_value[]=3&field_document_type_value[]=specification&speclib=

[mvitkun]

44 minutes ago, patrickjp93 said:

12 seconds of searching.

Click on any of those documents, it'll lead to this.

[patrickjp93]

3 minutes ago, mvitkun said:

Click on any of those documents, it'll lead to this.

So stop being a scrub and make a free account.

[79wjd]

11 minutes ago, patrickjp93 said:

So stop being a scrub and make a free account.

Then it's not 12 seconds of searching....unless you expect people to make an account any time any website asks for it just because of one stubborn asshole. 

 

by the way, I'm still waiting for that Denver K1 source, which you still haven't linked.... It's fine if you don't like providing sources, but then don't continue to make claims after other people have provided contradictory sources. It's okay to be wrong. 

[Dabombinable]

8 minutes ago, djdwosk97 said:

Then it's not 12 seconds of searching....unless you expect people to make an account any time any website asks for it just because of one stubborn asshole. 

 

by the way, I'm still waiting for that Denver K1 source, which you still haven't linked.... It's fine if you don't like providing sources, but then don't continue to make claims after other people have provided contradictory sources. It's okay to be wrong. 

I still remember him encouraging me to pirate a text book-which was US only and expensive as fuck. Instead of actually providing a source of information that everyone could peruse.

[patrickjp93]

13 minutes ago, djdwosk97 said:

Then it's not 12 seconds of searching....unless you expect people to make an account any time any website asks for it just because of one stubborn asshole. 

 

by the way, I'm still waiting for that Denver K1 source, which you still haven't linked.... It's fine if you don't like providing sources, but then don't continue to make claims after other people have provided contradictory sources. It's okay to be wrong. 

HMCC does the same thing. So does Intel. Get down in the muck and do the work yourself or quit complaining.

 

I provided a link directly from Nvidia's product brief pages, so you can go suck eggs. It's fine to be wrong, but I stick to the facts, so it's incredibly rare that I am.

 

6 minutes ago, Dabombinable said:

I still remember him encouraging me to pirate a text book-which was US only and expensive as fuck. Instead of actually providing a source of information that everyone could peruse.

Sometimes the luxury doesn't exist. Do the work yourself and prove my sources wrong.

[79wjd]

5 minutes ago, patrickjp93 said:

HMCC does the same thing. So does Intel. Get down in the muck and do the work yourself or quit complaining.

 

I provided a link directly from Nvidia's product brief pages, so you can go suck eggs. It's fine to be wrong, but I stick to the facts, so it's incredibly rare that I am.

 

Sometimes the luxury doesn't exist. Do the work yourself and prove my sources wrong.

Except that there isn't a single source showing that the Denver K1 is used in anything other than the Nexus 9. 

 

Also, the only source you linked says that Denver would be in the Nintendo NX. That is literally the only source I see that you linked. 

[i_build_nanosuits]

5 hours ago, patrickjp93 said:

http://www.guru3d.com/news-story/pci-express-4-will-deliver-75-watt-over-pcie-not-300w.html

 

The short of it? Tom's Hardware (and many others) got their analysis wrong. It's 75W per slot, same as always, and 75*4 = 300, just like I said when the spec launched.

 

 

 i knew this couldn't be true right from the beginning....just imagine how hot the slot providing 300W would be...it would probably melt.