Overclock.net - An Overclocking Community - Reply to Topic

Thread: [AMD] Official AMD PCIe statement and new driver 16.7.1 coming in 24 hours Reply to Thread
Title:
Message:

Register Now

In order to be able to post messages on the Overclock.net - An Overclocking Community forums, you must first register.
Please enter your desired user name, your email address and other required details in the form below.
User Name:
If you do not want to register, fill this field only and the name will be used as user name for your post.
Password
Please enter a password for your user account. Note that passwords are case-sensitive.
Password:
Confirm Password:
Email Address
Please enter a valid email address for yourself.
Email Address:

Log-in


  Additional Options
Miscellaneous Options

  Topic Review (Newest First)
07-07-2016 08:39 PM
Majin SSJ Eric
Quote:
Originally Posted by Ghoxt View Post

I hear you and even understand where you are coming from. Except at the same time, many of us own and actively use both AMD and Nvidia hardware and feel compelled obviously to let our opinions be heard regarding any discreet GPU matter as we see fit. And no, you didn't say otherwise, just pointing it out. biggrin.gif

What are we labelled? Do we even guess or know the percentage of us that have and still use both in our household? Steam won't help with that lol.

My opinion is that just as this was overblown like you said for something that likely would have affected a small percentage of RX480 owners and possibly near any normal failure rate. Fact is we won't know now of course as we don't have a time machine to test both scenarios, with the new adjusted Bios and without it.

Interspersed in many of the comments are even purely Nvidia owners that wholesale recognize this as overhyped obviously and is just one leg of the perception of the bad AMD PR/marketing table. It's not the end of the world of course.

All that said:
This was not something AMD is ignorant about. Knowing the expected testing that reviewers are going to perform. Knowing the Internet "Reaction" to any perceived problem in today's market, and finally Knowing the performance / electrical envelope of their own product, I'm more shocked that AMD did not see this coming.

I don't actually disagree with anything you said there. I myself have owned many cards from both companies and have had strong opinions both for and against them both. And you're right, knowing just how petty and meticulous reviewers can be when testing AMD products they should have made damn sure that there was NOTHING that could lead to even the possibility of bad press from the 480, yet here we are, just days after launch with some minor discrepancy about PCIE slot power draw being touted far and wide across the internet as some massive design flaw that somehow totally discounts all the performance and value advantages that this card has over every other card in its class. *Note that it took forever for the 970's 3.5GB issue to come to light after its release and I think its fairly obvious why that is (mind you I think that "issue" has been massively overblown as well but at least that was less an oversight and more false advertising, unlike this situation). The fact is that overdrawing the PCIE slot by a few volts/amps is not a problem for 99.99% of mobo's and that most components are designed with a built-in safety margin factored into the spec. Its not an issue in a real sense but it is an issue in a PR sense and its something that AMD should have been on top of before releasing a product to reviewers who have a documented history of more stringent expectations of perfection from AMD (I'm not calling them malicious, mind you, as the lackluster reputation that AMD has earned in some cases has lead to some of this bias in part).

The main thrust of my post was that we have a select few Nvidia fanboys (some of which who seem to have a borderline obsessive focus on every post I make) hyping this situation to the moon as though the 480 is some kind of dangerous virus that must be eradicated from existence when, in actual fact, it is really as close to a non-issue as you can get with literally only a handful of claimed problems being reported out of a customer base in the tens of thousands already (and those claims are not even verifiable in any way). I absolutely know that there are a lot of members here on OCN who regularly patronize both companies in the GPU sector and generally post very fair and honest critiques when warranted and was not denigrating them with my post...
07-07-2016 11:06 AM
cdawwgg
Quote:
Originally Posted by stargate125645 View Post

Ha, so I must have read that release a couple of times and my brain substituted 24 hours each time.

I read this in Zoidberg's voice. Well played Dr.
07-07-2016 10:00 AM
stargate125645
Quote:
Originally Posted by sugarhell View Post

Quote:
Originally Posted by stargate125645 View Post

It's been 24 hours...

The post says 48 hours but the title 24
Ha, so I must have read that release a couple of times and my brain substituted 24 hours each time.
07-07-2016 09:57 AM
sugarhell
Quote:
Originally Posted by stargate125645 View Post

It's been 24 hours...

The post says 48 hours but the title 24
07-07-2016 09:53 AM
stargate125645 It's been 24 hours...
07-07-2016 08:26 AM
KarathKasun
Quote:
Originally Posted by jomama22 View Post

I realize the use of power planes/layers for component use, as I said, aib (add in boards) are going to have some interference (whether caps, resistors or what have you) that do not let the 24/20 pin connectors have a direct connection to things such a sad pcie. This would only make is easier to diagnose issues stemming from the component itself as opposed to the PSU.

These parts would be damaged much earlier and prevent failure at the atx connector itself. That's the only point I'm trying to come across.

A blown atx pin is almost certainly not coming from an aib.

The 24 pin connecton directly powers the PCIe 12v plane. It is a simple 12v electrical bus on many motherboards. People have been melting the connector with 3/4 way Crossfire/SLI setups since that became a thing.
07-07-2016 04:53 AM
jomama22
Quote:
Originally Posted by Blameless View Post

It doesn't take much looking to find an example of almost any combination of components that appear to have destroyed one another.

I'm inclined to side with those that state the issue is overblown because if it wasn't, a far larger portion of boards used with the RX 480 would be experiencing issues.
Since forever. It's why PSU quality matters and why power supplies even have more than one output voltage.

The +12v and +3.3v lines from the ATX 24-pin are connected to power planes that directly feed the PCI-E slots, with maybe a few passive resistors and capacitors between them. You mess with either output voltage at the PSU and it's easily measured at the slot.

Only where 3.3v, 5v, or 12v are not suitable (modern CPUs, memory, chipsets, etc) do we have VRMs that sit between PSU output and components. PCI-E, PCI, USB, fan headers, and others typically make use of power planes that are directly supplied by the PSU.

In the past things were even more direct. Go back to the early/mid-90s and you have +3.3v and 5v from the PSU directly supplying processors, memory, chipsets, and PCI/ISA devices, with almost no voltage regulation or power conditioning between.

I realize the use of power planes/layers for component use, as I said, aib (add in boards) are going to have some interference (whether caps, resistors or what have you) that do not let the 24/20 pin connectors have a direct connection to things such a sad pcie. This would only make is easier to diagnose issues stemming from the component itself as opposed to the PSU.

These parts would be damaged much earlier and prevent failure at the atx connector itself. That's the only point I'm trying to come across.

A blown atx pin is almost certainly not coming from an aib.
07-07-2016 04:43 AM
Blameless
Quote:
Originally Posted by Defoler View Post

Point is that there had been several video reviews posted on OCN several times now of reviewers saying that some systems have been shutting down, some people posted pictures on reddit of their burned ATX connectors or their dead PCIE slots.

Some people dismiss it as fraud that it is no proof, yet they also have no proof that it is untrue as well.
So... give and take. You can't claim one thing is "overblown" when you go on something else and "overblown" that.

It doesn't take much looking to find an example of almost any combination of components that appear to have destroyed one another.

I'm inclined to side with those that state the issue is overblown because if it wasn't, a far larger portion of boards used with the RX 480 would be experiencing issues.
Quote:
Originally Posted by jomama22 View Post

Since when do we feed direct current from a 24 pin connector to any aib component? that doesn't even make sense.

Since forever. It's why PSU quality matters and why power supplies even have more than one output voltage.

The +12v and +3.3v lines from the ATX 24-pin are connected to power planes that directly feed the PCI-E slots, with maybe a few passive resistors and capacitors between them. You mess with either output voltage at the PSU and it's easily measured at the slot.

Only where 3.3v, 5v, or 12v are not suitable (modern CPUs, memory, chipsets, etc) do we have VRMs that sit between PSU output and components. PCI-E, PCI, USB, fan headers, and others typically make use of power planes that are directly supplied by the PSU.

In the past things were even more direct. Go back to the early/mid-90s and you have +3.3v and 5v from the PSU directly supplying processors, memory, chipsets, and PCI/ISA devices, with almost no voltage regulation or power conditioning between.
07-07-2016 03:48 AM
Ghoxt
Quote:
Originally Posted by jomama22 View Post

A major issue here is the fact that brand loyalty has such a profound impact on the outcome of such issues. I don't think anyone here can deny that nvidia owners have a much larger install base then AMD. Sale and profit numbers as well as ownership numbers make that very obvious.

You make a damned good point, Admittedly. Yup.

07-07-2016 03:46 AM
jomama22
Quote:
Originally Posted by Ghoxt View Post

I hear you and even understand where you are coming from. Except at the same time, many of us own and actively use both AMD and Nvidia hardware and feel compelled obviously to let our opinions be heard regarding any discreet GPU matter as we see fit. And no, you didn't say otherwise, just pointing it out. biggrin.gif

What are we labelled? Do we even guess or know the percentage of us that have and still use both in our household? Steam won't help with that lol.

My opinion is that just as this was overblown like you said for something that likely would have affected a small percentage of RX480 owners and possibly near any normal failure rate. Fact is we won't know now of course as we don't have a time machine to test both scenarios, with the new adjusted Bios and without it.

Interspersed in many of the comments are even purely Nvidia owners that wholesale recognize this as overhyped obviously and is just one leg of the perception of the bad AMD PR/marketing table. It's not the end of the world of course.

All that said:
This was not something AMD is ignorant about. Knowing the expected testing that reviewers are going to perform. Knowing the Internet "Reaction" to any perceived problem in today's market, and finally Knowing the performance / electrical envelope of their own product, I'm more shocked that AMD did not see this coming.

A very harsh "opinion" might see it as doing a disservice to their loyal customers allowing this to occur in the first place as they are not uninformed little 15 year olds. No different than the 970 crap NVidia pulled, they are paid professional GPU engineers and Architects so moreso from that fact some might not feel they deserve a pass in this debacle.

But then again, should we hold any of our GPU Overlord Masters to a higher standard? A slave wonders... redface.gif

A major issue here is the fact that brand loyalty has such a profound impact on the outcome of such issues. I don't think anyone here can deny that nvidia owners have a much larger install base then AMD. Sale and profit numbers as well as ownership numbers make that very obvious.

Because of this, any issue that arises will be treated completely different depending on the product developer.

I think it is fair to compare csgo users to GPU fanboys, at least from a vocal standpoint. Like almost anything, the louder crowd will always win the shouting match irregardless of the actual information.

We can even look at the 970 vram issue as a great indication of this. On this very website, we had threads ranging in the thousands of posts with proclamations of defense and offense against nvidia. What became of that? Nothing, literally nothing at all. We heard some lame reasons from nvidia and reviewers all but forgot and forgave then for the 3.5 GB of vram.

But God forbid amd has what...a ******* pcie overdraw(which every GPU in the mid-high range has done for years and years) and the sky starts to fall and amd is forced to release a driver that appeases these people by doing things those same users don't even understand.

It's very frustrating especially from someone who uses both amd and nvidia. The last thing any consumer needs is internal warfare among customers in a oligopoly segment. Yes, let's bash one of two companies still making competitive GPU products because you like the color/name/popular vote. It's so sad to see what this has all turned into.

I can only wish that nvidia becomes the sole supplier of dgpu's so that everyone must pay $500 for a 1360 in year 2020 so I get to have my 1380tis at $2000 a pop all to myself.

It's just so ridiculous, like what is there to gain from bashing one company or another over things so overblown?

I have no problem with calling out real, factual, informative problems when needs be. But the idea that crowd ideology has become the end all be all of information is just sad....really sad.

I'd like to hear from just one person, ONE, who has posted in this or the other 480 pcie thread that has actually encountered and blown pcie slot.

To the point of amd having to know their product would face pcie banter. Almost all gpus' exceeded the 75w limit whether in constant overdraw or sharp peaks. This is nothing new. There is nothing special that makes the 480 draw some exceedingly exuberant amount of current through the pcie. The only only reason this has become visible is through some really clever (and I mean that, clever) proclamations of motherboards being destroyed using the 480. None of which have been confirmed by any outside source other then the posters themselves.

I highly doubt amd just sat around and tested the 480 only only one mobo or didn't test it on an outright diagnostic rig that tells them every power draw of every pin (yes, every pin) that connects to their card. There is something fundamentally wrong with the assertions many people are making about this card. Hell. I'll put my money where my mouth is and buy $1000 worth of 480s and testing equipment just to prove otherwise.
This thread has more than 10 replies. Click here to review the whole thread.

Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off