AMD: Speculation, Rumors, and Discussion (Archive)

[DavidGraham]

95W was attained through overclocking.
You clearly didn't read or comprehend the article.

The comprehension problem is yours, not anyone else's. Any sustainable overdraw outside of spec is bound to cause problems, the rep didn't set the lower troublesome limit (after which problems begin to arise) at 95w, that would be ridiculous, baseless and completely random. He just quoted that number specifically because this is what X480 is drawing, OC'ed or not. The amount of overdraw can vary from application to application, could be 100w at stock, could be120w if OC'ed, and as the article says we still to see the absolute WORST case scenario of this overdraw.

What company takes the blame from homemade overclocking?
Is this a troll post?

Not talking about legal blame here, but a public one. Why should I buy a piece of silicon that violates spec when OC'ed when other silicons don't?

 

[xEx]

Also something some here are missing: the PCE slot is spec to give 66W through the 12V + the 3v that adds 75W. the RX is asking for 80 on the 12v alone. from 5.5A that the spec says to 6.5 sustain.

 

[BRiT]

lol this is going to be viral

This is not going to help out anyone. If AMD can't recover from this, expect to be paying $1500 for top graphic cards next, and $400 for entry-level graphic cards.

 

[3dilettante]

This is not going to help out anyone. If AMD can't recover from this, expect to be paying $1500 for top graphic cards next, and $400 for entry-level graphic cards.

If Nvidia wanted a good zinger, they could give a price chart for their Founders Edition GPUs with the price of a new motherboard listed as savings.

 

[pharma]

This is not going to help out anyone. If AMD can't recover from this, expect to be paying $1500 for top graphic cards next, and $400 for entry-level graphic cards.

It's unfortunate, but AMD at this moment seems like a pretty shoddy company where image is more important than true technical achievement ....

 

[xEx]

This is not going to help out anyone. If AMD can't recover from this, expect to be paying $1500 for top graphic cards next, and $400 for entry-level graphic cards.

I am actually an AMD fan and I actually really dislike Nvidia and Intel as a company believe it or not lol. But I have to be objective.

 

[xEx]

It's unfortunate, but AMD at this moment seems like a pretty shoddy company where image is more important than true technical achievement ....

Well...it...is right now for them.

 

[pharma]

Well...it...is right now for them.

You are correct ...

 

[Razor1]

It's unfortunate, but AMD at this moment seems like a pretty shoddy company where image is more important than true technical achievement ....

The problem is they probably thought they could get away with it for what ever reasons and they really needed to push it cause otherwise they will obviously get killed in reviews if they had less performance or more power consumption......

As I stated before they really needed to hit these reviews out of the ball park, its a lot of pressure.

 

[A1xLLcqAgt0qc2RyMz0y]

Just to raise a separate point,custom AIB will not overcome this problem by implementing an auxiliary PEG 8-pin/2x6-pin/etc configuration unless they also re-define the power distribution and ease preassure off the PCI Express slot, this is going to be critical with AIB partners as it is more than likely they will at a minimum increase the power targets and possibly some small OC.

CHeers

They must "re-define the power distribution and ease preassure off the PCI Express slot" or this problem (systems shutting down, killing PCI-E slots) will only get worse.

The question is how easy is that to do and how quickly can they do it.

 

[Geeforcer]

I wonder if we finally found out exactly what validation the chip cryptically-allegedly failed at above 800MHz.

 

[A1xLLcqAgt0qc2RyMz0y]

The problem is they probably thought they could get away with it for what ever reasons and they really needed to push it cause otherwise they will obviously get killed in reviews if they had less performance or more power consumption......

As I stated before they really needed to these reviews out of the ball park, its a lot of pressure.

It really is looking like AMD pushed the clocks too high to compete against the clocks Nvidia was getting from Pascal realizing that the 1060 would also have high clocks and along with the rumors of GF 14nm process problems resulted in what we see the 480 being today.

 

[xEx]

It really is looking like AMD pushed the clocks too high to compete against the clocks Nvidia was getting from Pascal realizing that the 1060 would also have high clocks and along with the rumors of GF 14nm process problems resulted in what we see the 480 being today.

The tdp is actually OK, the clock vs volt Is broken but the tdp is ok. The problem is on the power delivery.

Enviado desde mi HTC One mediante Tapatalk

 

[CSI PC]

It really is looking like AMD pushed the clocks too high to compete against the clocks Nvidia was getting from Pascal realizing that the 1060 would also have high clocks and along with the rumors of GF 14nm process problems resulted in what we see the 480 being today.

It is not necessarily the clocks per se that are the problem.
It is the power target/voltage envelope they applied by default.
You can get pretty close performance to defaults just by lowering the voltage values and in turn reduces the power consumption.

Question that will never be answered is why the default settings and what were they focused on for their expected ideal parameters.
I would still recommend this card over a 970, just reduce the voltage a bit and no to increasing clocks/power target.
Cheers

 

[3dilettante]

It really is looking like AMD pushed the clocks too high to compete against the clocks Nvidia was getting from Pascal realizing that the 1060 would also have high clocks and along with the rumors of GF 14nm process problems resulted in what we see the 480 being today.

There's also the "minimum VR spec" lower limit that AMD itself placed at the 290X/GTX 970 level.
Even without the 1060, how far down does it seem AMD had to go before it misses its own mark?

Some of the averages don't show much breathing room.

Interestingly, that was from Roy Taylor, whose being unleashed was what I cited elsewhere as my go-to for determining where AMD's "hoisted by your own petard" point would be.

 

[ninelven]

If it is so, someone should be fired...

I'm not so sure. Really, many, many people within AMD should have known better all the way up to the top. That no one spoke up and stopped this from happening I think is indicative of the larger cultural issue(s) I mentioned earlier. So unless the person doing the firing is also willing to fire themselves, I don't think it should be done. And even if you fire somebody what good does it do? I am sure (or at least hope) that everyone involved feels terrible about this happening (and I bet it doesn't happen again..). Now, if a person directly responsible genuinely doesn't realize that they screwed up or is somehow still trying to defend or hold on to the board design being the right decision, then yeah maybe that person needs to go. Because that may be indicative of a source of the cultural problem(s) I think may exist at AMD...

If AMD can't recover from this

While it is certainly not a good thing, I don't see it being a huge long-term problem for AMD. The chip itself is still quite decent, it just needs a better board. AMD can even treat this as an opportunity to build trust and relationships with their consumers by making sure they feel safe and satisfied it whatever the ultimate outcome may be. But please god don't let them pat themselves on the back on social media for doing the right thing in a @#%$ situation (this is something I would not be surprised to see from the current/past culture).

 

[Razor1]

If AMD does a quick response and fixes the issue expediently there won't be any real financial impact on them, if they let it go or give this problem more time to fester, yeah its going to be a real big problem.

Either way they have mud on their face right now, its how they clean it up is what matters, either proactively fixing it or not.

 

[Alexko]

I remember the days when overclockers had to fiddle with jumpers, pencil mod their CPUs, lapped their dies, installed coolers that could and would take parts of their fingers off, and did so with the knowledge they could very readily turn part of their rig into magic smoke.

I still remember when I polished a heatsink with sandpaper because the original finish was terrible, or when I removed a couple of crappy little fans from a Leadtek GF4 Ti 4600 to replace them with a 92mm fan attached, not to the card but next to it, to a bent PCI slot bracket, with pieces of string and adhesive paste.

Good times.

 

[flopper]

If AMD does a quick response and fixes the issue expediently there won't be any real financial impact on them, if they let it go or give this problem more time to fester, yeah its going to be a real big problem.

Either way they have mud on their face right now, its how they clean it up is what matters, either proactively fixing it or not.

Maybe a drunk engineer pushed the wrong tangent when bios was coded, it can happen and that they stated oc dreams of the fury and we know how that turned out.
so many engineers on the staff and none saw this coming?

 

[Razor1]

Thing like this don't go unseen, for what ever reason they did it anyways. There are many checks and balances even before getting PCI-SIG certification, there would be no way it would have been overlooked.