Separate names with a comma.
Discussion in 'AMD Flavor' started by KickAssCop, Sep 10, 2015.
One image to summarize.
*R9 Nano Fair Review Thread
Thank you for making this. I have read Ryan's review (pcper) so far, and will read all of them throughout today.
A few points I will make, based on Ryan's review so far.
First, I do not see the dynamic clock speed as an issue. In our own paper launch article I spelled out that we have to think about how the dynamic clock speed works, and to not think of it as throttling. The Nano doesn't "throttle." Instead, it dynamically keeps the clock speed in a range locked to the TDP. So far, the testing I have seen, supports that fact. The range Ryan got, is what was reported to us by AMD. AMD quoted to me the range should be 800-950MHz. This is correct according to his testing. Therefore, the clock speed range is not in question.
A sound issue
Ryan detected a coil whine issue - http://www.pcper.com/reviews/Graphi...w/Noise-Testing-SFF-System-Build-Temperatures
The Case Size Issue
My question, and issue, is the fact that Fury X can fit in some of these cases AMD is targeting for Nano. There are mATX and ITX cases that can fit full-size video cards.
The Fury X is a small video card itself however, just like the Nano, granted you need room for the rad/fan, but an mATX case will have that space, and most ITX cases will as well, cramped yes, will it work? Yes.
I have seen ITX builds on youtube with Fury X working inside. I have seen mATX and ITX builds with full-size cards as well.
In Ryan's own review, he admits the case he used for testing supports a Fury X and a full size PSU.
So my question is, why Nano when you can Fury X in the same case?
If your answer is power, I remind you that case supports a full-size PSU, and others might as well. Even if it didn't, you could still probably find a powerful small PSU.
Price Price Price
Since Nano is the same price as Fury X, and you can put Fury X in most of these cases (I admit there are some that it probably won't fit in) then the price does not make sense. The Nano should be a bit cheaper.
One thing so far I have not seen talked about is that Fury X at $649 is no where near 980 Ti performance, form factor aside. Therefore, Nano is even less powerful to 980 Ti, form factor aside. With the Nano being even farther apart from 980 Ti in performance, speaking strictly from a performance basis, form factor aside, it makes even less sense at the price it is asking.
Finally, yes the small 970 is slower, but it is also a LOT cheaper. On a price vs. price comparison alone, they are not comparable.
I will be reading more reviews now.
Sits right between a 390X and the non-X Fury. $230 and $120 more respectively. Uses less power, is cooler, but also louder.
I don't get Tweaktown's conclusion that this makes the Fury and Fury-X pointless.
The Nano only makes sense if you absolutely MUST have a case that doesn't support a full size GPU, meaning you don't care about wasting money.
If your case fits a full size GPU, the Fury X, the Fury, the 980Ti the 980, and the 390X all make much more sense for the money.
I predict they will make 10 of these for the 40 total people in the world that would even consider trying to use it..
AMD said "look what we can do" and built an okay product that almost nobody needs or wants.
The fact that is also seems to have a coil whine just seems unfathomable considering the kinds of cases this is made for would typically be so small they are limited on sound dampening.
Even with unlimited budget of someone elses money I would never even consider this card.
Here are my observations.
1) I am impressed by the performance out of this card. I expected it to perform worse than a 980 so given that, I think packing this much performance in such a small form factor is a feat in itself.
2) The card is noisy as sin. It is making noise higher than a 980 Ti with custom cooling and beating or matching 980 Ti reference cooling (which is also noisy as fuck all). For a quiet small PC, it is a complete deal breaker and mentioned in every review I have read.
3) The price is ridiculous but I think that horse has been beaten to death. It is such a small market that anyone who buys this doesn't have money in their minds as a concern. I am not one of those people so for me this card is pointless to begin with.
4) The card throttles. Whether you call it dynamic or whatever, I am never going to see the 1000 MHz advertised speed unless I overclock or set speed in AB but even then I am not guaranteed the 1000 MHz.
5) Minor nitpick but no backplate? Seriously? Fury X gets it for 649 and Nano doesn't. lol
AMD PR was unnecessarily idiotic is my observation.
My verdict. Decent card for custom PC builders. For everyone else, they need their brains checked before they shell out the cash. If the fan were silent, I would even say that the 649 tag MAY be justified in a nice small box next to my 4K TV. Oh wai... that only has HDMI input.
The no HDMI 2.0 part is a very important fact considering the market this card is aiming at. More so than Fury X or Fury.
Granted there is a supposed adapter for DP to make that happen, but I haven't seen one yet.
The problem there is really a marketing issue, AMD markets the clock speed wrong. Instead of an "Up To" clock AMD should state a Base clock like NVIDIA. In that case, anything above the base clock would be considered a boost clock.
For example, saying the Nano has an 800MHz base clock, then in testing you get 825MHz and upwards, you would see it as and think of it as boosting beyond the base clock, just like NVIDIA GPU Boost. NV was smart when it created GPU Boost, it is the perfect way to think about the clock speed. AMD's description is always going to make you think it is throttling under the advertised speed.
The KING of small factor builds.
Best card performance in the small factor market.
When is the Hardocp review out? oh you guys didnt do fair reviews and Kyle has beef with AMD guys, sry forgot. But I dont care about what Kyle or Brent thinks anyhow or what reviews they do anyhow and that is why I write this to just clarify I dont care about that at all.
The Small Form Factor Market has decided to build cases that fit full size cards.
AMD has built a Tech Demo that they are offering up for sale. It only fits the niche of SFF builds that don't support full sized cards, in the niche of people that want SFF cases.
From Anandtech's review
My response: Then it shouldn't be the same price.
That case could easily house a Fury X or larger card.
So, now that we know it performs very well, it's the size (which was already known) that is wrong...
This is not 2002 where small form factor cases were shite and didn't fit any real video cards. This is 2015. Plenty of good SFF cases have been made that accommodate full length cards and even WC loops. With no HDMI 2.0, tremendous amount of noise for a SFF HTPC type build and the stupid price - what was AMD thinking? What was the market for this?
One of the major flaws I see with these reviews is they all test the card in a rather large case (for a SSF).
This brings two issues:
#1. It opens up comparisons to full size cards, which makes the Nano look to be significantly overpriced.
#2. It does not properly test the "target" for the Nano. If it is intended to truly function in tiny cases where a full size card would not fit, I want it tested in such a small space to see how it handles heat. I think you might end up seeing some serious throttling if you actually put it in a case where you need something that small.
Guru mde sure to use a really hot case. And it showed it beating a 980 quite easily which is neat. So not sure how these oc'ed 970s are going to do but 980 is pretty lackluster if there's any heat.
What i find interesting is how much power they can save with dynamically clocking the card, without losing too much performance.
Voltage decreases as well, with clock, hence the power drop.
Looked at a few of the reviews..
Really bad coil whine is still not enough of a factor to not give it a high "award"???
All they had to do to get rid of the coil whine was better coils or stick some dampening glue on the coils. At the price point this thing should have had absolutely no coil whine.
One thing I will say is that even my 7970s have a bit of coil whine, but only when running 3dmark in a few lower tests that get 1600+ fps. I cannot imagine having a system that makes that noise while gaming.
Oh.. and the vrm temps on the back of the board????
Wow....this card is a beast in some titles. Nano needs a water AIO STAT !!!!!
I can see why it's so expensive, AMD must have culled some of the best Fury chips for the R9 Nano.
THIS THING OVERCLOCKS??!!
Well as bunch of reviews, and the verdict is in.
Generally slower than a 980 and faster than a 970.
100 bucks more than a 980 and 300 more than a 970.
And judging by the Fury X market we can expect the price to jump 100 to 150 if it sells as well as the Fury X.
That being said AMD has some serious balls releasing a tech demo/proof of concept card into the retail channel.
I don't trust the Guru review one bit. They never once mention what case they used for testing and the only picture of the card in case they have up is massive.
If that's what they consider the target case for this card they are out of their minds. That thing would fit a full size card easily and get much better performance for the dollar.
I stand by my statement of not seeing single review where the form factor of the card was necessary.
It performs very well for a $430 card...but they are charging $650 for it, so it is the price that is the problem.
The mention of SFF cases fitting full sized cards is merely pointing out that the small size of the card is basically not a relevant excuse for spending $650 on a card that performs no better than a $430 card.
I agree. I know the PCper guys seemed confused as to the purpose of this card on the first podcast where they previewed it.
I know some people like small cases, but we have so many options that can fit a normal sized card. You can't praise them for building something we do not want or need. I am tired of hearing how it is amazing they put so much power in such a small card because it doesn't matter.
My one request for the potential future [H] review would be that they throw it in a case small enough to not be able to fit a full size card or even a Fury X. I'm really curious to see how it handles heat and power draw in the form factor it's aimed at.
See, I was really looking forward to this card. I want to build a tiny PC about the size of a gaming console to stick in my entertainment center. I had hoped this card could fit that build well as it would give me the size/power I want to get great image quality out of a small form factor. AMD really botched this launch and it seems they picked review sites who weren't willing to test the thermal limits in such a small enclosure.
You haven't seen TomsHardware review then. Toms' put the Nano in a closed case for core clocks and temperature testing.
The first thing that comes to mind for me is a refreshed MacPRO with Dual Fiji cards. Seems like the perfect fit! Price can pretty much be whatever, and a big missed marketing opportunity lost for AMD.
Nice little card, it creates a lot of build possibilities. The price... Well it's not a price of consider buying no matter what the performance is. And frankly if supply is as limited as the Fury X, I'd expect the price to climb even higher.
#1. Tom's shows significant clock drops in the closed case
#2. The comparison benchmark data was taken from the open air test bench run, not the closed care real world.
I'm not sure why people are having such a difficult time understanding that performance inside a closed case that is heated up can differ from an open air test bench.
What makes me laugh is the editors choice award, despite them admitting it's overpriced, has loud coils, loud fan and can't even do the 4k/60hz living room that a SFF PC like this is targeting.
Is it small and relatively fast? Sure. It fails every other metric, though.
Fixed it for ya
Faster than 980 when in a hot case..... Read the reviews
So if you raise the powertune on this card, I assume it will stick to 1000mhz as long as you can cool it, np?
generally, at 1080p, it is on par or slower than a 980, and in most cases that are higher resolution than 1080p, it is faster than the 980 by a good 10% or so (with a few exceptions) .
Seriously, if this card was priced at 980 (or within 5-10% of it) I would have picked one up for FreeSync, but what's the point of saving $200 on a Freesync vs Gsync Monitor, only to spend it on a GPU that offers me the same performance?
So it is a decent card that is priced out of its market segment like everyone has been arguing? Color me surprised . Would probably be a good buy if it was priced around $450 US, small nagging issues aside.
I don't know the answer to your question, but if you had the room to cool it, what would be the point of having it? It only makes sense if it is in a case so small you can't use anything else in which case your cooling ability is going to be seriously limited
it seems like the card packs quite a wallop for such a small size... but dat price doe
if AMD dropped prices on the Fury line by 50 - 100 i'd be all over it, but NOOOOOOO
Yeah pretty unfortunate they limited their potential customer base with the price. I'm sure a lot of people would be all over this thing if it was around the ~$500 mark. Now only people who must have a card this size and want most of the power will be picking this card.
They already have that. It's called the Fury X. It costs the same and offers better performance. The Fury X even fits in most high end enthusiast SFF cases. SFF gaming evolved to incorporate 500w+ PSUs and full sized GPUs while reducing total volume. This product is a year late
I wanted this to be so much more competitive. I'm waiting for AMDs Zen/DDR4 offerings before my next upgrade and this has me looking to green and blue more and more.
So has anyone found a review where it was actually tested against an OC'd 970-ITX? That's the only question I see mattering