AMD Radeon RX 580 8GB Review

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.
It'd be interesting if Tom's started including undervolting results in their reviews. I'm curious to see if the RX 580/570s tend to undervolt as well as their RX 400 counterparts (and what effect that has on power consumption), or if jacking up the clock speeds ate away any undervolting room.
 


Skewed toward what? The relative performance I gathered from this review seems pretty similar to that performance summary at Techpowerup. Obviously the exact numbers are going to vary based on what games and settings they choose to test, as well as what hardware the test system uses, and which versions of the other cards they test them against.

That review may have used results from cards at their reference clocks, while Tom's used factory overclocked cards for this review to provide a more realistic representation of how the cards compare to others currently available for sale. AMD didn't offer a reference design for the RX 580, so they're testing a factory overclocked model, so it makes sense to place it up against other factory overclocked models.
 
It's awful 220Watts for a mid range card whereas the 1060 consumes 120W there's a 100W difference this is not 50 as many say. Again I still don't know if I should buy this one or the 1060 DAMN! i'm exhausted of reading reviews and forums I need to make up my mind.
 
It's awful 220Watts for a mid range card whereas the 1060 consumes 120W there's a 100W difference this is not 50 as many say. Again I still don't know if I should buy this one or the 1060 DAMN! i'm exhausted of reading reviews and forums I need to make up my mind.
 
Same issues with this review as many of them. The RX series has been known to hold better MINIMUM frame rates. Why should we even care whether or not it's 60 fps or 900? If frames dip below 60fps then we should start questioning a card's value. "Average" frames are not a good quantifier of a card's abilities anymore.
 
Cons

New name may confuse some
Click to expand...

You mean like the GTX 1060 3GB and 1060 6GB? SOME will confuse the two, thinking it is the exact same chip without any differences other than RAM amount. You never made a major point about that (you did mention it though.) Before you go off, yes, I know it is the same Gen with the 1060 vs different gens with the 480 vs 580. NVidia has done this in the past, AMD has as well, with a fairly long span between new tech. Intel has done it too in recent history. It adds a smell when you pointedly and repeatedly point it out for one manufacturer and with others it gets mentioned once, and then maybe only in a footnote.
 
Can I compare any of the benchmarks with CS:GO? I´m thinking about building a Ryzen 5 1600 and rx480 system, but I almost only play CS and don´t know what to expect from a system like this??
 
Can I compare any of the benchmarks with CS:GO? I´m thinking about building a Ryzen 5 1600 and rx480 system, but I almost only play CS and don´t know what to expect from a system like this??
 
You get all what you need tu build your own, individual conclusion: average and min FPS, FPS curves, frametimes and variances as charts/curves and an unique uneveness index. For each game. The text around the charts isn't the holy bible and each reader can take from this menu, whatever he personally prefer or believe. :)

And to be honest - 60 min FPS may be worse than 45 FPS with better frametime variances. This is the reason why we are using so many charts graphics to compare it in detail.

 


my impression is with the RX400 series AMD want to tell the world that they have improved their power efficiency with GCN architecture. remember polaris 11 demo at CES 2016? the main focus is power efficiency instead of performance vs competitor product. with RX500 AMD is like "screw efficiency and lets push this chip to it's limit".
 


Guess I must be in the 2%. I don't mind the extra data points as long as the stock clock performance is also included. I've seen some reviews which only show fully overclocked components and those are totally worthless to me.
 
Vega was suppose to be out last winter but AMD kept pushing back the release date. It should've been out by now. In my opinion, AMD wants to get as much profit out of the Polaris chip as they can before releasing Vega. In 2014, Nvidia's Maxwell caught AMD with their pants down. So AMD harried Polaris to catch up but Polaris was simply to little too late. Shame. Hopefully, Vega will be the future arrived.
 


Polaris was never anything but rumored to be out as winter. It was hopeful thinking on people looking a the extremely vague road-map they put out. AMD did however push the release date from Q1 2017 to Q2 2017 and it's widely believed this was due to HBM2 availability.
 


Vega was never supposed to be out last winter, in fact all people had was a vague point on a graph with very wide spaces between the years, that folks then extrapolated Q1 2017 from. Clearly that hasn't happened yet.

For some reason people take every rumor about AMD and turn it into fact and then get pissed off when it doesn't happen. Happened quite a bit with Ryzen, the hype machine went ballistic and everyone was expecting a 1800X that demolished every processor Intel sells for half the price. AMD never said or alluded to that, the hype machine drove that. Instead we got the exact processor anyone who paid attention to the official releases would have expected. And its still a great processor. But the hype machine is saying that AMD failed.
 


AFAIK Vega have always been 2017 product. but then you have some desperate people wanting to see AMD beating nvidia they start making their own speculation based on tiny info they can found on the internet. for example the rough roadmap that AMD show to the public. they play around with that image and based on the position AMD put Vega on the chart was and they make conclusion that Vega is 2016 product. that october rumor there are even some people treat it like pure fact. at another forum there are some people that try holding people back from buying 1080/1070 saying that the card is hard to find and you might not be able to get one until july time frame after limited initial batch for launch. they said it is a bad decision to get 1080/1070 in july when three months later Vega will come to the market. of course there is also leak about AMD investor slide that mention Vega will be 2017 product. even mention 1H 2017 release meaning it is not even confirmed to coming out in Q1 2017. but as usual some people only want to believe want they want to believe.

anyway AMD did not rush polaris to compete with nvidia. like it or not AMD have still to wait until it is feasible for them to launch polaris because it all depends on how well the new node are doing. but starting with polaris AMD change their marketing strategy. usually they will not tell us much until the product is officially launch or very close to launch. but starting with polaris AMD start the "hype" 5 to 6 months in advance before launching the product. to be honest i'm not sure if i like the idea. AMD need to remind people that they still have new product coming out but at the same time they also giving away some info to their competitor. polaris 11 vs GTX950? nvidia know AMD try to get polaris 11 under 75w so nvidia tweak GTX950 so the power consumption will go down from 90w to 75w but doing so without negatively affecting initial 950 performance.
 
 
Two things I have to say about this review:

1) Let's be honest, is Civilization VI really a good game for testing GPUs? I mean, it's probably GREAT for testing CPUs but it seems to me that it would be far more CPU-limited than GPU-limited since there isn't much eye-candy.

2) "We like the extra performance, but could do without AMD's proclivity for re-branding existing products. "

I can see Tom's Hardware's point on this, although at least there is a measurable difference between the cards. What I don't understand is why it's only a bad thing when AMD does it. Nine years ago, Chris Angelini DEFENDED nVidia for rebranding the G92 chip for the SECOND TIME! First they rebranded the 8800 series as the 9800 series, then they rebranded the 9800GTX+ as the GTS 250! You know, some of us have been around long enough to remember:
http://www.tomshardware.com/reviews/geforce-gts-250,2172-11.html

Or maybe six years ago in a review from Don Woligroski where he acknowledged that the GeForce GTX 560 was just an overclocked GTX 460:
http://www.tomshardware.com/reviews/geforce-gtx-560-amp-edition-gtx-560-directcu-ii-top,2944.html
In his conclusion however, he also refrained from bashing nVidia for the refresh/rebrand. In fact, he specifically stated that even though the GTX 560 used more power than the Radeon 6950, few would notice or care. (Another thing that the Radeon card got bashed for in this review.). I personally thought that Don's review was the best of the four because he was extremely professional about it with no attempt to promote the card(s) involved or to bash them. He simply presented the facts, said a few things that were neutral and true and that was that:
http://www.tomshardware.com/reviews/geforce-gtx-560-amp-edition-gtx-560-directcu-ii-top,2944-17.html

I think that it's rather unfair and hypocritical of Tom's Hardware to do this. I know that the reviews are by different people but there's no consistency except a seeming willingness to forgive nVidia where AMD would be chastised. Now, it's very possible that this isn't the case because Chris didn't attack AMD for rebranding the HD 7970 to the R9-280X but I have NEVER seen nVidia chastised for doing this and they've done it FAR MORE in their history than AMD ever did. Even to those who don't agree with me, I'm sure you can see where I'm coming from.

Don was right you know, nobody would notice or care about the so-called "vast differences" in power use between cards because it would take years to even begin to scratch your wallet. Take it from me, I use an R9 Fury for Christ's sake, a card with a TDP of 275W that can approach 400W under heavy load but still stays pretty damn cool.

The fact remains that we have Chris and Don defending nVidia for what amounts to be very little differences in their rebranding but we have Igor taking AMD to task when the difference between their rebranded card and original card was actually something worth talking about. And then making DAMN SURE to focus on power use which is a pathetic excuse for taking a shot at ANY piece of computer equipment. Fermi had by far the worst performance/watt of any card architecture I'd ever seen or heard about but nobody here had an unkind word to say about the company itself like Igor did. You guys are professionals, you should be better than this. I don't mind if you want to take a company to task for something that they deserve but what's good for the gander is also good for the goose and you should be doing it ALL the time, not just when it's a Radeon card.
 

I'm not Chris or Don and if you know my reviews - I'm not biased. I also wrote the same about other companies if I was thinking, it is the time to say what I see and mean (and it is a translation, the original is a lot smoother). I do not prefer brand A or B, but quality, innovation and truth. This is, what I missed with the RX 580 this time. It isn't a new card and it isn't a new chip. But it was sold from AMD as something and special new. I like Polaris and if you read my Radeon Pro WX 7100 review, you can also read, where the sweet spot is and what I'm thinking about the 480. I talked with Mike Mantor and Raja personally and this chip was designed as, a pure 110-watts-GPU, nothing else. To increase the voltage up to 1.25V and to say, hey - here is a new product with higher clocks, isn't they way to impress me. It's simply a lie, nothing else. And it is not important for me, who does it, AMD, Nvidia, Intel or someone else. Read my RX 480 roundup and you will see, which cards I prefer. Not the fastest or most expensive, but the cards with the best engineering and craftmanship inside. It needs a lot of to impress me, but the fastest way is truth and the respect for the own product. And here AMD simply failed. The media is not so stupid to believe all. Nothing else.

 
Does anyone know of an actual RX580 roundup?? Similar to the RX480 roundup they did......? There's ~20 options for the new RX580 & comparing them has been a PITA...
 


Don't know of any roundups yet, but I always recommend getting an AMD card from Sapphire, ASUS or XFX. Whichever has the best price. You can somewhat ignore the factory OC'd models as they are usually minor OC's that you can do yourself, without paying the price premium. It really comes down to a good, quiet cooler.
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom