• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

The thread which sometimes talks about RDNA2

Status
Not open for further replies.
GDDR6 will give you error corrections (so lower fps) if overclocked over 2150 or so. In fact you should check what is the frequency that gives you the most fps ( between 2000-2150Mhz ).
Yep, 2012Mhz show’s performance improvement but it drops after that. Core clocking is where it’s at.
 
If there was a memory bottleneck, oveclocking the memory would lead to gains, but it’s core overocking that scales well from my testing.

GDDR6 will give you error corrections (so lower fps) if overclocked over 2150 or so. In fact you should check what is the frequency that gives you the most fps ( between 2000-2150Mhz ).

The memory is clocked quite high from stock. But then if it is bottlenecked a core overclock wouldn't provide any performance gains.

Hardware unboxed review of the sapphire Nitro the stock OC had a greater performance uplift at 4k than at 1440p. I'm off to play games now i will check some more reviews later.
 
Those tears are delicious.
Since you can't actually name what amd did other then price higher with low availability. I wouldn't call it scummy.

But sell all available stock to miners making people wait weeks/month. Along with all the other things mention and still defend them. You ate righ. It takes one to ignore all that.
Rolf
:D

"Sell all available stock"... so everyone with FE cards in their sig is just making it up?
 
GDDR6 will give you error corrections (so lower fps) if overclocked over 2150 or so. In fact you should check what is the frequency that gives you the most fps ( between 2000-2150Mhz ).

Think the OC reviews I saw the VRAM was always maxed out at 2140, never any higher. Guess that’s probably why then.
 
Think the OC reviews I saw the VRAM was always maxed out at 2140, never any higher. Guess that’s probably why then.
Yes i think the 2138 is the max the card is running but even with a modified bios, if you go over that you will lose fps.
And everyone should test what is the frequency where he gets the most fps at 4k rendering (2000-2150) because the cards are not the same.
 
A user from reddit has down some power limit testing with 3rd party software

seems the big Navi cards are either overvolted from stock or AMD has the overclocked the cards hard to make them competitive cause just like Ampere, Bug Navi retains most of its performance with significant power savings

you can maintain 95% of performance with a 20%/60w reduction in power draw

and you can maintain 75% of performance with a 50%/150w reduction in power draw

They're not clocked quite as hard as ampere though but still over their efficient peak, on the rtx3090 you can save 80w/22% and maintain 90% performance so the 3090 is overclocked from factory slightly harder than the 6800xt but both cards are clearly overclocked pas their peak efficiency and you can achieve significant power reduction for just 5-10% performance loss

What was the software? Igorlab stuff? I'm interested in getting the 6800XT but want to undervolt it.
 
Watch the HU video and steve said that AMD told him they will have AIB at MSRP in around 4-8 weeks. That's around 40 days that I gueesed at before...ROFL
That's why I said I would just wait it out.

Ampere released on Sept. 17, 2020 and it's be about 73 days since released. About 2 months and even Steve admits there has been no signs of improvement in stock for Ampere.

IMO, AMD should put more priority of RDNA2 over Zen 3 since Intel won't have anything anytime soon.
 
Last edited:
It is pure speculation - if you actually go to the source of the information they only speculate that a certain amount came from cryptocurrency and then further down compare the mining performance of Ampere - the original report itself doesn't tie that income to Ampere only to crypto. It is the news articles that have reported on that which have tied the two together.

I have no doubt they (Mining entity) will have muscled in early on the scene as it offers them access to much needed factors. The hardware out there was getting stagnant and Ampere/RDNA2 was a step up in performance for no price hike. They would easy have tapped into the supply chain especially in eastern countries where backhanders would be common. I guess it would perfectly explain why there is pressure on the retailers as the low volume manufactured was not able to spin out of the mess.
 
Watch the HU video and steve said that AMD told him they will have AIB at MSRP in around 4-8 weeks. That's around 40 days that I gueesed at before...ROFL
That's why I said I would just wait it out.

Ampere released on Sept. 17, 2020 and it's be about 73 days since released. About 2 months and even Steve admits there has been no signs of improvement in stock for Ampere.

IMO, AMD should put more priority of RDNA2 over Zen 3 since Intel won't have anything anytime soon.
though isnt zen 3 more profitable than the gpu series, thats what i read, could be wrong though
 
though isnt zen 3 more profitable than the gpu series, thats what i read, could be wrong though

The Zen 3 server chips I would have thought would be higher priority than GPU, it's AMD's main battleground where they can really make a killing and push out Intels monopoly
 
though isnt zen 3 more profitable than the gpu series, thats what i read, could be wrong though
I think per wafer they are more profitable but I do agree with the priority thing from a business stand point.
On the cpu front if you are waiting for a 5000 series CPU you are unlikely to go Intel. While there is definitely a dwindling market to address on the gpu front.

There is also the need to increase GPU market share which has other benefits down the line
 
I have no doubt they (Mining entity) will have muscled in early on the scene as it offers them access to much needed factors. The hardware out there was getting stagnant and Ampere/RDNA2 was a step up in performance for no price hike. They would easy have tapped into the supply chain especially in eastern countries where backhanders would be common. I guess it would perfectly explain why there is pressure on the retailers as the low volume manufactured was not able to spin out of the mess.

That is probably true - but a different story to nVidia selling massive amounts of Ampere directly to miners.

I know there have been a few reports of corruption with the supply chain in some parts of the world - Russia and parts of Asia, etc.

There are only 3 guys with multiple accounts. One has a 3090, one a 3080 and you with the 3070.

So lack of a serious answer just makes my point for me.

though isnt zen 3 more profitable than the gpu series, thats what i read, could be wrong though

Zen will be vastly more profitable - when you are whacking out wafers of 70-80mm2 parts and able to charge $200-300+ with yields near 90% versus 500+mm2 and yields well below 90% you need to stick some significant markup on those GPUs to make sense commercially.

TBH if AMD wasn't a GPU company and/or depending how much volume they can sell Zen wise they probably wouldn't be able to justify producing those GPU wafers at all.
 
Last edited:
The Zen 3 server chips I would have thought would be higher priority than GPU, it's AMD's main battleground where they can really make a killing and push out Intels monopoly

Plus getting traction in the server space can take some time due to hardware lifecycles and ROI's that businesses want to achieve before refreshing their equipment. So AMD need to stay on it while they are in front to make sure they are converting all the businesses they can to the red team.
 
Status
Not open for further replies.
Back
Top Bottom