• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Chiphell: AMD 3xx perf. leak

People introduce that issue by overclocking, they also have the tools to fix the issue. If you want any company to spoon feed you high overclocks without tweaks, you are going to be disappointed.

Wrong I had this issue at stock, I almost sent my card back because of it. Show me the tools to fix this issue. I want to see them.

I will be waiting patiently for these tools so my spoon fed ass can fix the problem without using a modded bios. And I am sure many people on OCN would love to see these tools as well.
 
Wrong I had this issue at stock, I almost sent my card back because of it. Show me the tools to fix this issue. I want to see them.

I will be waiting patiently for these tools so my spoon fed ass can fix the problem without using a modded bios. And I am sure many people on OCN would love to see these tools as well.

Your card was faulty then. I've had three 970s and none crashed at low loads in either of two PCs.
 
Your card was faulty then. I've had three 970s and none crashed at low loads in either of two PCs.

The card isn't faulty it was a problem with Nvidia BOOST bins and voltage tables with lower ASIC cards. The problem is fixed permanently my flashing to a modded bios and changing some of the boost states and voltage in the voltage table.

It was particularly more prevalent as soon as you overclocked yes, but if you had a G1 with a boost clock of 1400mhz out of the box you were going to have this problem. Like I said it only ever happens in certain menus of games in low utilization scenarios such as Mordor, BF4 and The Evil Within to name a few I had issues with. Their is a post about how to " bake in" your clocks to get rid of the problem as well as using the voltage table mod to fix it.

But yes modded specific boost state bins completely fixes this issue now that we have mod capabilities. It was not so nice before that however.
 
Last edited:
Even if it is an overclocking problem, how is it not a problem with the card?
 
Sadly, this AMD card wont see its maximum potential on an AMD gaming rig without PCIE3.

AMD 990fx is limited to PCIE 2.0

FM2+ chipsets might have PCIE3.0 but lacks in CPU raw power.

AMD fucked themselves into a corner on their own chipsets.
 
Sadly, this AMD card wont see its maximum potential on an AMD gaming rig without PCIE3.

AMD 990fx is limited to PCIE 2.0

FM2+ chipsets might have PCIE3.0 but lacks in CPU raw power.

AMD fucked themselves into a corner on their own chipsets.

Heh I think they know that by now. But it isn't going to stop them from continuing to do their best in the desktop graphics arena.
 
Sadly, this AMD card wont see its maximum potential on an AMD gaming rig without PCIE3.

AMD 990fx is limited to PCIE 2.0

FM2+ chipsets might have PCIE3.0 but lacks in CPU raw power.

AMD fucked themselves into a corner on their own chipsets.
Didn't they announce they weren't going to expand their FX line like 6+ months ago?
It sounds like they're just pushing for APU's and discrete GPU's at this point.

Or did they announce some new FX chips coming up? I honestly haven't bothered to keep up with AMD CPU's since, uhh... 2006.

I'd like to see them scrap the entire CPU department and move all their funds into GPU's but I guess they would argue the PS4/XBO and mobile devices are more important.
 
Even if it is an overclocking problem, how is it not a problem with the card?

It's not an overclocking problem as an overclock has no control of lower power states and the boost clocks and the bins associated with those overclocks. It IS an under-clock and under-volt issue. Basically when Nvidia BOOST finds a low utilization point(pre determind boost bin) it under-clocks the gpu and voltage to match the "bin" ( basically a pre-determinied voltage and clock state that is considered stable by Nvidia). But the 'bin' on the 970s when reaching a certain utilization would "cross bins". AKA the clocks would be a couple bins higher than the voltage rating for those bins, causing instability and crashes.

This has yet to be fixed by Nvida or even acknowledged.

It is a major problem and it affects all 970 brands. And the ONLY long term fix is to flash to a modded bios that has fixes to the voltage table to the "bins" that match the the boost state and it's Pstate.
 
Last edited:
Didn't they announce they weren't going to expand their FX line like 6+ months ago?
It sounds like they're just pushing for APU's and discrete GPU's at this point.

Or did they announce some new FX chips coming up? I honestly haven't bothered to keep up with AMD CPU's since, uhh... 2006.

I'd like to see them scrap the entire CPU department and move all their funds into GPU's but I guess they would argue the PS4/XBO and mobile devices are more important.

If they scrap the CPU dept, Intel would have a monopoly on CPU price.

You'll be playing on your 6 year old I7 wishing you could afford a $4000 new CPU.
they wouldn't lock down the price if there was no competition.
Certainly it looks like there is no competition as far as performance, but imagine if there was absolutely none at all.

AMD needs to stop fucking around and start research into and deliver a NEW top-end CPU. Even if it's within 5-10% performance of a top-end Intel CPU.
 
If they scrap the CPU dept, Intel would have a monopoly on CPU price.

You'll be playing on your 6 year old I7 wishing you could afford a $4000 new CPU.
they wouldn't lock down the price if there was no competition.
Certainly it looks like there is no competition as far as performance, but imagine if there was absolutely none at all.

AMD needs to stop fucking around and start research into and deliver a NEW top-end CPU. Even if it's within 5-10% performance of a top-end Intel CPU.

on the desktop side they pretty much already have a monopoly.. AMD's pretty much given up on it at this point and completely shifted to the mobile platform side which they are pretty far ahead of Intel on. if you look at their roadmap, they have no plans on releasing a new consumer high end desktop processors any time soon. and to be honest i don't think they ever really tried to compete with intel after they lost the dual core performance war when Intel released the core 2 duo architure until their attempt with bulldozer and flopped, they just filled a market intel wasn't filling at an affordable price point.

the GPU side on the other hand Nvidia and AMD have been throwing punches back and forth for the last 6 years and it's been interesting to watch how each company tries to find a way to compete with the other on which ever architecture they release. and also seeing how each company switches from brutforce performance to efficiency then back to brutforce performance and so on through each architecture.
 
Sadly, this AMD card wont see its maximum potential on an AMD gaming rig without PCIE3.

AMD 990fx is limited to PCIE 2.0

FM2+ chipsets might have PCIE3.0 but lacks in CPU raw power.

AMD fucked themselves into a corner on their own chipsets.

Except for the fact that PCI-E 2.0 isn't really a limiting factor.
 
Except for the fact that PCI-E 2.0 isn't really a limiting factor.


AMD's CPU's/APU's being a limiting factor of their GPU's, now that is something that could be quite real factor. It's kind of sad and embarrassing at the same time when you think AMD's own products may be holding it back from their own products potential lol.

They really need to get their x86 shit together or find someone to merge with.
 
AMD's CPU's/APU's being a limiting factor of their GPU's, now that is something that could be quite real factor. It's kind of sad and embarrassing at the same time when you think AMD's own products may be holding it back from their own products potential lol.

They really need to get their x86 shit together or find someone to merge with.

Most people have known for a years that AMD CPUs can hold the full potential of the GPU back and the only thing embarrassing at this time is that its a new revelation to you.
 
AMD's CPU's/APU's being a limiting factor of their GPU's, now that is something that could be quite real factor. It's kind of sad and embarrassing at the same time when you think AMD's own products may be holding it back from their own products potential lol.

They really need to get their x86 shit together or find someone to merge with.

Have you been asleep for the last 8 years?
 
AMD's CPU's/APU's being a limiting factor of their GPU's, now that is something that could be quite real factor. It's kind of sad and embarrassing at the same time when you think AMD's own products may be holding it back from their own products potential lol.

They really need to get their x86 shit together or find someone to merge with.

sure, if you want to consider terribly coded game engines developed by companies that still think dual core processors are the norm for every gamer then yeah sure their CPU's could be holding back their cards performance. but hey what ever floats your boat..
 
sure, if you want to consider terribly coded game engines developed by companies that still think dual core processors are the norm for every gamer then yeah sure their CPU's could be holding back their cards performance. but hey what ever floats your boat..

So ?

Do you live in perfectly coded utopia or in reality ? Because if it happens to be reallity then there is no place for AMD cpu in any non low-end gaming rig.
 
you really think a card with a 4096 bit bus is going to work properly on pcie 2.0?

umm....

why would the internal Vram memory bus width have any bearing on PCIe bandwidth?
 
Last edited:
sure, if you want to consider terribly coded game engines developed by companies that still think dual core processors are the norm for every gamer then yeah sure their CPU's could be holding back their cards performance. but hey what ever floats your boat..

Go show us how it's done, cowboy! Solve inherent computational and design capabilities for the world! Thanks in advance!
 
umm...I was agreeing that PCIE 2.0 isn't a limiting factor. Are you saying it is or isn't?
I'm saying that pci-e 2 is not a limiting factor. My bad, I went bad and re-read your statement and the context of it. I misread the context and thought you were saying that you were tired of people saying pcie2 is not a limiting factor.
 
Go show us how it's done, cowboy! Solve inherent computational and design capabilities for the world! Thanks in advance!

Well, he is right, considering the amount of computing power we have now, there is far less emphasis on efficiency in the programming world, add insanely competitive landscape, where companies are pushing games out faster than a chicken can lay eggs....
 
Well, he is right, considering the amount of computing power we have now, there is far less emphasis on efficiency in the programming world, add insanely competitive landscape, where companies are pushing games out faster than a chicken can lay eggs....

Which doesn't change anything. No matter what is cause the end effect is the same - CPU with 4 as strong as possible threads is optimal choice for today.
 
Which doesn't change anything. No matter what is cause the end effect is the same - CPU with 4 as strong as possible threads is optimal choice for today.

I'm not arguing what is best, I'm just agreeing with lack of efficiency :) since he asked him to do a better job, indicating that the best job was already being done.
 
Well, he is right, considering the amount of computing power we have now, there is far less emphasis on efficiency in the programming world, add insanely competitive landscape, where companies are pushing games out faster than a chicken can lay eggs....

far less? Its non existent.

Nothing like the Nasa guys used to do to reduce lines of code to achieve the same result. Naturally as hard-drives and ram evolved the strict requirements for space have been erased.

I mean the space shuttle has like 424kb of memory. So all emphasis was on efficiency then.

Nowadays its like will it run? yes? will it crash? no, ok good to go. doesn't care if it uses 30gb.
 
I'm not arguing what is best, I'm just agreeing with lack of efficiency :) since he asked him to do a better job, indicating that the best job was already being done.

Games are inherently difficult to multithread. Go do a little googling even if you aren't familiar with game programming and loops and you'll very readily understand my comments. Long story short, not using 8 or 16 or more cores heavily has nothing to do with developers being "lazy" or "poor optimization" or game release schedules :p.
 
actually,its DX11 and prior thats difficult to multithread, not games themselves.

Mantle for example is very easy to multithread (and Dx12 by extention)
 
actually,its DX11 and prior thats difficult to multithread, not games themselves.

Mantle for example is very easy to multithread (and Dx12 by extention)

Completely and utterly incorrect. Any programmer in existence that isn't on hello world can tell you that, including myself.
 
Games are inherently difficult to multithread. Go do a little googling even if you aren't familiar with game programming and loops and you'll very readily understand my comments. Long story short, not using 8 or 16 or more cores heavily has nothing to do with developers being "lazy" or "poor optimization" or game release schedules :p.

And you're speaking from experience?
 
Completely and utterly incorrect. Any programmer in existence that isn't on hello world can tell you that, including myself.


Clearly youve never made it past hello world.

I suggest you google some game programmer logs.

The threading bottleneck is DX11.

Period.

Its not even up for debate as its been discussed to dealth
 
Last edited:
Clearly youve never made it past hello world.

I suggest you google some game programmer logs.

The threading bottleneck is DX11.

Period.

Its not even up for debate as its been discussed to dealth

I am a programmer, FYI. But it is basic parallelism concepts that fundamentally and inherently cause game loops to be notoriously difficult to substantially multithread. The operating system or api have no relevance or relationship to this fact. This is common knowledge, no debate needed indeed, just as sure as water is wet. Thank you ;).

Here's one readily available developer statement from the first page of Google results about this that I told you to go look at earlier rather than continuing to spread misinformation :

I think there are opportunities to split RIFT’s work into several threads and really exercise modern gaming computers. The programmers reading this are cringing in sympathy right now, as adding support for multiple cores is often a nightmare, but it’s worth trying for the framerate boost that might be achievable.
http://forums.riftgame.com/general-.../352460-tech-talk-rift-performance-2-2-a.html

Again it isn't poor optimization or lazy devs as you falsely claim.
 
Last edited:
I just bought my son a GTX970, and I really like my GTX780 but a 390x will be very tempting if it does come on 16nm. Man, 2015 may be an expensive year. My 5930k will be here hopefully early January :D
 
I just bought my son a GTX970, and I really like my GTX780 but a 390x will be very tempting if it does come on 16nm. Man, 2015 may be an expensive year. My 5930k will be here hopefully early January :D

We won't be seeing high end 16nm gpus until the end of 2015 in all likelihood or even the start of 2016, from either camp. You may see a shrink of an existing chip like GM204 but that's about it. 28nm is here to stay for awhile yet!
 
Back
Top