AMD Acquired HiAlgo. Bye bye stuttering.

[Trixanity]

8 minutes ago, don_svetlio said:

Exactly - DigitalFoundry and Science Studio show it beating the 970 and in SS' case it ran at 78*C at reasonable fan levels. For some reason engineering samples are wildly inconsistent.

Well, Polaris did introduce on-boot "calibration" to optimize your card according to the chip quality, so in theory not all card performs the same whereas cards would normally follow the lowest common denominator and otherwise rely on boost clocks or overclocks to show case the difference in chip quality.

 

Anyway, that's all theory. Oh and given the new node, I'm guessing chip quality has a fairly wide range compared to the very optimized 28 nm.

[don_svetlio]

22 minutes ago, Bouzoo said:

I've seen that it runs in load at 80+ C and I think that's more than realistic, since that is wildly consistent to all that I've seen, from sites like Guru3D, TPUP and Tom's HW. And the noise is... well nothing worse than a 290X or 980Ti from their tests.

Depends but yeah. In any case, it's not terrible. But we expected more.

[AlTech]

1 minute ago, don_svetlio said:

Depends but yeah. In any case, it's not terrible. But we expected more.

In my case I expected way too much and was basically asking for unrealistic things.

[Bouzoo]

2 minutes ago, AluminiumTech said:

In my case I expected way too much and was basically asking for unrealistic things.

Wait for AIB cards, I think they'll be what we expected from the reference card. Expect maybe for the power draw, if that bothers you.

[AlTech]

1 minute ago, Bouzoo said:

Wait for AIB cards, I think they'll be what we expected from the reference card. Expect maybe for the power draw, if that bothers you.

No. I was expecting more than  the craziest OC on the planet could offer.

 

I was expecting Fury or Fury X performance.

[Bouzoo]

1 minute ago, AluminiumTech said:

No. I was expecting more than  the craziest OC on the planet could offer.

 

I was expecting Fury or Fury X performance.

Well you were expecting too much. Hands down too much.

[AlTech]

1 minute ago, Bouzoo said:

Well you were expecting too much. Hands down too much.

I did warn you.

 

In my case I expected way too much and was basically asking for unrealistic things

 

So it looks like RX 490, unless they manage to screw that up as well.

[don_svetlio]

4 minutes ago, AluminiumTech said:

No. I was expecting more than  the craziest OC on the planet could offer.

 

I was expecting Fury or Fury X performance.

 

3 minutes ago, Bouzoo said:

Well you were expecting too much. Hands down too much.

Well, at leas tthe 480 scales well with core OCs so a Fury-like OC might still be possible

[KillDog]

The RX 480 has already managed to eliminate stutter completely thanks to the async compute engines. The frametimes with it are much better than the 970 or 980. 

IDK why many people don't see the potential of this card, they only see benchmarks and then reply it's a shit card.

[Prysin]

1 hour ago, Colonel_Gerdauf said:

I really want to know what you have been reading, but most reviews (video or not) report it performing slightly worse than the 970 in quite a number of theoretical and real-world scenarios.

it ended up worse then the 970 in NVIDIA "THE WAY ITS MEANT TO BE PLAYED" titles.

[Doobeedoo]

That's awesome. This will really be a great improvement. Cause stutter can really affect online play so it can be a serious issue.

[AlTech]

9 hours ago, KillDog said:

The RX 480 has already managed to eliminate stutter completely thanks to the async compute engines. The frametimes with it are much better than the 970 or 980. 

IDK why many people don't see the potential of this card, they only see benchmarks and then reply it's a shit card.

What do you mean? All GCN GPUs feature Async compute Engines. And yet some may have stuttering.

 

I never said it was a shit card. I said I was disappointed. There's a huge difference between the two.

[themaniac]

10 hours ago, don_svetlio said:

 

Well, at leas tthe 480 scales well with core OCs so a Fury-like OC might still be possible

dont forget that the 480 is still in the very early drivers stages so more optimized drivers allong with a great oc and it might get to fury like performance

[Khvarrioiren]

5 hours ago, themaniac said:

dont forget that the 480 is still in the very early drivers stages so more optimized drivers allong with a great oc and it might get to fury like performance

The biggest problem the RX 480 faces is the power delivery and cooling. AIB solutions will most likely aid the performance a whole lot more than drivers atleast initially.

[Prysin]

17 hours ago, Trixanity said:

Well, Polaris did introduce on-boot "calibration" to optimize your card according to the chip quality, so in theory not all card performs the same whereas cards would normally follow the lowest common denominator and otherwise rely on boost clocks or overclocks to show case the difference in chip quality.

 

Anyway, that's all theory. Oh and given the new node, I'm guessing chip quality has a fairly wide range compared to the very optimized 28 nm.

no. Bristol Ridge which launched back at the E3 event Introduced on-Boot calibration.

[AlwaysFSX]

Won't that create a delay in inputs?

[RagnarokDel]

That was a pretty nice interview

[handymanshandle]

20 hours ago, themaniac said:

does AMD/RTG really have the money to acquire, even if their really small

It could pay off for them.
Also AMD/RTG can afford this acquisition, I mean RTG is a division of AMD. So if RTG proposed the acquisition, every executive on that division probably had to agree with it and then AMD's execs probably had to agree with it. So I think that them acquiring HiAglo is fine. I think they're also banking on the RX 480 being a good "fundraising" card for them. Which I think it will be once they fix the PCIe spec issue. Which sounds like may also affect aftermarket cards.

[RagnarokDel]

Global Foundry is known for their chips not liking heat, and apparently AIB were reporting that the lottery is strong but some samples can reach 1600mhz on air.

[handymanshandle]

12 minutes ago, Prysin said:

no. Bristol Ridge which launched back at the E3 event Introduced on-Boot calibration.

Wasn't Bristol announced/paper launched at Computex? All AMD said at the PC Gaming Show at E3 was that the RX 480, RX 470 and RX 460 would be the entire line of Polaris cards this year, and no word on Vega's ETA. I'm not sure if Vega is still slotted for early 2017. Since rumors have said sometime this fall, likely October. Which I don't believe since AMD has otherwise been pretty quiet about Vega. Besides a couple tweets by Raja a couple weeks ago or whatever where he and the team that worked on developing Vega celebrated something or another. 

2 minutes ago, RagnarokDel said:

Global Foundry is known for their chips not liking heat, and apparently AIB were reporting that the lottery is strong but some samples can reach 1600mhz on air.

Doesn't AMD have priority with Samsung, or is that only for vram?  

[Prysin]

@AlwaysFSX

in what way?

 

@wcreek

Bristol Ridge is a OEM SKU foremost. So we wont see it in laptops before back-to-school sales start.

And my bad, yes it was computex.

[handymanshandle]

1 minute ago, Prysin said:

 

 

@wcreek

Bristol Ridge is a OEM SKU foremost. So we wont see it in laptops before back-to-school sales start.

And my bad, yes it was computex.

Yeah it's intended for laptops and it's built off of pile driver or excavator iirc.

 

Here's a little bit of something unrelated...

Yeah I wouldn't mind getting a laptop with an A9 or A12 off Bristol Ridge but being able to do a couple things that I wanted to do in the next 9 or so months is going to be interesting to achieve, looking at it from an realist/pessimistic view it's unlikely it'll happen to do both of them. But from an optimist view, it's possible I could make both work.

 

[Guest Kloaked]

20 hours ago, Thony said:

My gtx 760 stutters before it reaches its vram limit. So I dont agree with your theory.

Don't bother with them.

[AlwaysFSX]

26 minutes ago, Prysin said:

@AlwaysFSX

in what way?

Depending on when the game decides to take inputs before rendering the frame it might cause input lag. Say at the end of the previous frame instead of when drawing the next frame. I know it's an issue in some engines.

[Bouzoo]

21 hours ago, Thony said:

My gtx 760 stutters before it reaches its vram limit. So I dont agree with your theory.

I would say that (probably) he means the VRAM cap limit massive stutter. It's been recorded in many scenarios how it stutters like a power point presentation when it hits 3.5-3.7 GB limit, it became evident about the time Dying Light came out iirc, with the key word being massive, not "regular" stutter.