Introduction
AMD’s new series of graphics cards, the RX 2XX series, is split up into the R9 Enthusiast class and the R7 mainstream class. Both segmentations feature impressive levels of value for money and top end performance at each price point but today we are looking at one of the higher end models. We have with us the Sapphire AMD Radeon R9 280X Vapor-X OC Graphics card in this review but we’ve also reviewed the AMD R7 260X and AMD R9 270X for today’s launch. While this is a Sapphire branded card we are using this as an opportunity to to also look at AMD’s Radeon R9 280X as AMD were not able to provide us with a reference design. The reference design of course is identical to Sapphire’s model other than the cooler is different and Sapphire have added a mild factory overclock taking the clock speed up from 1GHz to 1.07GHz, and the memory from 6GHz to 6.2GHz.
The Radeon R9 280X card we received from Sapphire and AMD is simply an OEM graphics card so it comes with nothing other than the card which is pictured above. Below you can see a GPU-Z screenshot of the new AMD graphics card which reveals those overclocks I mentioned earlier. A couple of other things to point out are that this card supports Direct X 11.2, OpenGL 4.3 and AMD’s new “Mantle” technology.
Looking at the AMD R9 280X more broadly it should have the following specifications:
Of course for those who don’t know, the AMD Radeon R9 280X is based on an optimised version of the AMD HD 7970 GPU. It features increased clock speeds and uses a more refined 28nm process as well as a redesigned cooling solution for the reference cards. The main advantage is that AMD is shipping the new R9 280X with a price of $299 compared to the HD 7970 which when it first shipped had an MSRP of $549. With the R9 280X, AMD hopes to take on Nvidia by offering a card that is faster than the GTX 760 and on par with the GTX 770 but costs $299 instead of the $249 of the GTX 760 or $399 of the GTX 770. Of course Nvidia will undoubtedly respond with price cuts but as things stand the R9 280X is seeking to be the GTX 770 killer by offering equivalent performance for a staggering $100 less. Let us proceed with this review and see if the R9 280X delivers.
Why don’t you use battlefield in the bench?
There is no Battlefield 3 benchmark. We’ve yet to find a decent benchmark tool that replicates consistent fair results.
Hi Ryan, I have one of these cards and am looking at getting a water block for it, do you know which AMD reference design would suit best? I would assume the 7970 design but would like confirmation if possible.
Excellent review Ryan (y)
I am not sure I would classify 166W slightly more power usage when compared to a GTX 770. That’s the equivalent of 2 processors and the sound is horrible. I personally would sacrifice the couple of fps just because of the sound and power usage. I guess for the price point it can’t be beaten but my other point still stands… Will wait and see what the 290x is like, hopefully the power and acoustic won’t be so terrible!
I think the excess power consumption is more to do with Sapphire’s board design than the R9 280X more broadly as I’ve stated in the review. From what others achieved in testing peak power should be only around 7970 GHz levels, maybe add 10-20W if its an OC card. It certainly shouldn’t be as high as it was from our testing
Yeah that’s fair enough that it is down to Sapphire’s design but the power usage is still horrible I think they need to go back to the drawing board with that. A shame really as Sapphire are supposed to be the best AMD card designers :s
no no no.. sapphire is the cheapest faulty stuff.. anyway, i new why i didnt pick 7970 two years ago, and i wont pick it now! lets wait for hawaii and see if they actually made some optimizations, so that the chip doesnt feel so heavy unlike nvidia, which does everything with grace and lightness..
Are you kidding me? What a troll
Calm your panties, the test is obviously broken.
Nr. 1 rule of looking up reviews: Don’t just blindly and solely follow the first one you come across…
The test is not “broken” in any way. TechPowerUp also showed that the MSI TF R9 280X consumes 65W more than a GTX 780 under Furmark load. That is pretty much exactly what our result shows. Ours shows 80W more than the GTX 780 with the extra 15W accounted for by the higher clock speed of the Sapphire Vapor-X model vs the MSI one TechPowerUp tested.
http://tpucdn.com/reviews/MSI/R9_280X_Gaming/images/power_maximum.gif
Number 1 rule of commenting on reviews – Trying to be a smartass gets you nowhere.
Or you could not be a patronising twat and read the other two comments in the thread before commenting…
If I have the Seasonic S12II 620W PSU, ASUS Maximus V Hero, 16GB Ram, and an Overclocked i7 4770K 4.2GHz, will the card be ok on my system?
Yes it should be. We were running a SB-E 3960X, X79 motherboard, couple of SSDs and this card and we only racked up 466 Watts so if you consider our PSU operates at about 85% efficiency that means about 400W Load on the PSU so that would be 220W spare for you. Plus you use Haswell and a relatively light motherboard so you’d have no trouble at all. It does depend on how many other drives you have connected and whether you use high voltages to overclock but if you keep this card close to 1100MHz you should be perfectly fine.
Wich 7870 did you used on the test the Tahiti or the Pitcaim?
The original 7870 so Pitcairn.
Hi Ryan, first off thanks for the awesome review as I was looking into the R9 280X which sells at 279€ here compared to GTX 770 which sells at 370€. Your findings in terms of power/noise to have me worried as my PSU is 550W from Antec.
(42A on the 12V rail)
My CPU is a bit less power hungry than the one used in the test so I would end up at 450W under furmark load based on your measurements. I checked other reviews where they tested power at gaming/GPGPU and while it still confirms eteknix findings that R9 280X is a power hungry beast they do show that during gaming the power consumption is more between a GTX 770 and 780 where it should be. Could it be that the R9 280X reacts specifically bad to Furmark type applications? Reference: http://www.tomshardware.com/reviews/radeon-r9-280x-r9-270x-r7-260x,3635-18.html
Thanks in advance for your feedback.
One additional question is if by any chance you measured the dimensions of the R9 280X Sapphire Vapor-X?
Are they compliant to the specsheet found here: http://videocardz.com/46415/sapphires-radeon-r7-r9-full-lineup-unveiled
Reason for asking is because my case can only house a 11 inch or 280mm card.
Hi there. Okay so first addressing the length. I’ve got the ruler out and we’re talking 27.5cm from the PCI bracket start to the longest point on the card. So yes it will just squeeze in and as the power connectors are top mounted there are no issues there. Of course the reference design is smaller so it depends what each board partner does.
Second with regards to power if you look at TechPowerUp they show similar results to us with regards to furmark http://www.techpowerup.com/reviews/MSI/R9_280X_Gaming/24.html but yes most definitely this extreme power consumption is largely to do with the way AMD cards react to Furmark. Under typical gaming loads you will find this consumes less than a GTX 780 or a 7970 GHZ edition. So on that basis 550W should suffice, but I would recommend not overclocking. Also what model of PSU is it? that will be important.
Thank you very much for the detailed response.
Length should be fine then and I am considering the Sapphire version you reviewed. It is in store at 279€.
(While GTX 770 is still being sold at 370€)
The PSU is a Antec truepower 550W with 42A (total) on the 12V rail.
My CPU however is a i5 750 2.66GHz which will consume about 20W less compared to the i7 used in your test setup.
From my understanding of the power requirements that should be fine then. 42 Amps is a lot for a 12 volt rail.
Thanks!
if the power consumption is basicaly the same as a 7970 then it can run on a good 550 psu right?
http://www.xbitlabs.com/articles/cases/display/chieftec-psus.html#sect1
Calin from what understand you should be fine given that your PSU brand is decent and that the amps on your 12V rail are sufficient.
well.50 A on 12 v rails are enough i guess
for the 280 x vapor x?
Ryan, an additional question related to the power measurement. Could you specify which furmark profile you used?
the Burn-In Test. default settings. Its an unlimited benchmark but we run it for only 5 minutes. (not full screen, AA off and 720p possibly 1080p can’t 100% remember will double check next time I fire up the test bench)
http://www.ozone3d.net/public/jegx/201204/furmark-1.10.0-main-interface.jpg
Thanks for the review! Is the voltages unlocked on this card like the Vapor X’s in the past? Thanks
I’m Afraid I don’t know. The HD 7970 Vapor-X was a mixed bag, some cards shipped with unlocked voltages others didn’t. I’d imagine this is the same, our sample was locked so I guess its luck of the draw but based on my experience I’d say they are locked.
Ryan even more questions 🙂
1) You mentioned possible compatibility issues with the motherboard due to the backplate. Would I have this problem on a P7P55D LE?
http://www.google.be/imgres?imgurl=http://www.ixbt.com/mainboard/asus/p7p55d-le/board.jpg&imgrefurl=http://ixbtlabs.com/articles3/mainboard/asus-p7p55d-le-i55p-p1.html&h=1103&w=1500&sz=547&tbnid=VLb-5sozJo1H7M:&tbnh=90&tbnw=122&zoom=1&usg=__Ue548EdBhyhpHLFknxlMniLe3DQ=&docid=4P2yTf49iFAl3M&sa=X&ei=xlVaUrrlFqeS0AXinoGYCA&ved=0CEQQ9QEwAg
I assume I can just plug it in in the second PCEX16_2?
2) Kitguru: “There is a little button on this card which enhances the core clock speeds by 20mhz (from 1,050mhz to 1070mhz) and memory speeds by 50mhz (from 1,500mhz to 1550mhz). The image above shows the switch ‘enabled’ – it glows blue.”
Can you confirm this overclock switch and if it is activated by default?
Reason for asking is that I want to burn-in the card using the furmark profile you used but in default mode.
The compatibility issue is with motherboard heatsinks, that board has no heatsinks so no compatibility issues will exist. I didn’t press the button, just ran default so I guess that means our card ran at 1070mhz out of the box.
Thanks a lot for consistently responding to questions in such a fast and detailed manner! Top notch sir!
No problem. That’s what we are here for. Thanks for visiting our site and reading our content.
Ryan I can’t seem to run out of questions. 🙂 When using Furmark in combination with GPU-Z it provides a GPU power draw value. Have you ever checked during reviews how accurate this value is compared to your multimeter measurements for power consumption?
Actually I haven’t. I’ve never really trusted values provided by software – it is normally inaccurate and underestimates actual power consumption.
For example if you review power supplies based on HWMonitor or OCCT voltages these voltages are very different to what you’d see on a PSU testing Chroma machine (which is more accurate). The same applies to Graphics Cards. GPU-Z offers a good “ballpark figure” but nothing concrete IMO.
Ryan, thanks for the great review. Been looking at the 280X the last two days comparing to the Nvidia 760, 770 and the 670 that just recently failed me. The 280X seems to be money for value and was looking at the 7970 to see if I can save money, but they haven’t been discounted here in Aus yet. But I prefer to go with a newer model even if it is a rebadge. I’m not fussed on the noise factor as when I game I have headphones on and loud and prefer the temp to be low as possible. Saying that, I was looking at this particular card, but in the AMD world, what GFX card manufacturers are to be more reliable. I’ve preferred Nvidia in the past, but now going for a AMD, I would like to know what brands make more reliable cards. The site I order from stock PowerColor, HIS, Gigabyte, Asus and MSI. Although there’s a difference up to $150 between the brands, what brand would your recommendation be in general and why?
Thanks again for the great review and it’s good to see you respond to peoples comments with well informed information or being honest and saying you don’t know instead of just winging a answer to a question.
Gigabyte and ASUS.
Reason why those two brands is because they offer the best performance with rather quiet fans with good warranty (RMA). MSI was rather carefull with their overclock so there you will have to do it yourself. HIS and Powercolor are not sold in my region but the brands are mentioned are usually complimented for using quality components.
Of those brands MSI and Asus are best. Both have a reputation for superior cooling and stability. Gigabyte is #3 (their OC cards are great performers but their reliability isn’t as high as Sapphire/MSI/Asus) while Powercolor/HIS are solid although unspectacular brands. They mostly sell stock cards clocked no higher than reference models.
i have personal bad experiences with msi and powercolor. msi cards have issues with going outside the voltages specified by original gpu manufacturers, and after a few years many of them needs overvolting to run stable, also their coolers can be exceptionally loud when worn by my experience. powercolor probably has the worst customer service of all aib partners in the world if you arent in usa.
if i were you, id go for the sapphire 280x toxic edition. its barely more expensive than other 280x’es, its the fastest of them all for now (often it ends up competing with gtx 780’s, at pretty much half the price), its runs at the lowest temps, it has some overclocking headroom, and while not having used sapphire myself yet ive only ever heard positive things about them. (actually, im likely to get one of them myself)
i have personal bad experiences with msi and powercolor. msi cards have issues with going outside the voltages specified by original gpu manufacturers, and after a few years many of them needs overvolting to run stable, also their coolers can be exceptionally loud when worn by my experience. powercolor probably has the worst customer service of all aib partners in the world if you arent in usa.
if i were you, id go for the sapphire 280x toxic edition. its barely more expensive than other 280x’es, its the fastest of them all for now (often it ends up competing with gtx 780’s, at pretty much half the price), its runs at the lowest temps, it has some overclocking headroom, and while not having used sapphire myself yet ive only ever heard positive things about them. (actually, im likely to get one of them myself)
I agree that Powercolor has awful customer service and that Sapphire is the best brand overall for AMD cards. But I was simply ranking the specific brands he specified the vendor carries. Personally I’ve never had problems with MSI cards.
Yeah everyone.. TAKE it from this guy! Because he knows what he’s talking about! ..PFFTTT….
The Toxic was rated the best r9280x / 7970 ghz out there… The above comment is one mans opinion and I would not take it to heart. Besides. You should be buying as low of speed as possible on this card and then OC it yourself.
Anyone can move a slider….
I’d have a number of recommendations for the AMD RX 2XX cards.
ASUS – generally the most power efficient and are normally very quiet and cool. Only downside is they often cost a fair bit more.
Gigabyte – often have the coolest running cards but at the expense of more noise than rivals (but still less than stock) They are often the cheapest but you sacrifice a little on the power efficiency and aesthetics (often use blue PCBs etc)
MSI – generally the quietest (TwinFrozr) but not always the coolest. Again they lack the advanced efficiency of the ASUS power componentry but look great, are quiet and often cheaper than ASUS.
Aside from those things you’d also want to consider warranty, whether it looks good enough for you and whether it fits into your system. Generally my personal preference is dictated by price, I have used ASUS DC2 cards and MSI TF cards in the past when they were on special. If I had to put my money where my mouth is I would buy an ASUS R9 280X Direct CU II (no OC or TOP as I’d OC myself manually).
HIS, XFX, PowerColor and others I have little experience with. Sapphire too I lack experience with, I have only ever used this card which I was informed wasn’t up to their usual standard as this was only an early production sample.
Hi Ryan, just tested the Sapphire R9 280X Vapor-X with custom/adjusted fan settings on 550W PSU.
Below tests were performed with the OC dialed back to 1050MHz core clock and 1500MHz memory clock.
In BF3 with fans at 45% it runs quite and stays at 60°C above 60% with all settings ULTRA. In furmark 1280×1024 resolution the fans were set manually to 100% and temperature maxed out at ~70°C (73°C). Similar result could be achieved with fans set at 60% which is less painfull on the ears. The 550W PSU gave no sign of issues in Furmark(15 minutes) while GPU-Z reported a GPU draw of maximum 220W (unreliable figure).
During Furmark I did notice that once the temperatures hit close to 70°C that the boost clock began to switch between 1050MHz and 950Mhz (reference core clock unboosted). Is this behavior normal? I assume this is the throttle down of the boost clock because certain key parameters like temperature are too high? (I hope it wasn’t power that was too low but then I would have expected this to happen as well at lower temperatures and it didn’t manifest below 60°C.)
Would love to hear your thoughts on the above Ryan.
The above 60%for BF3 needs to be above 60FPS (with high in the 90FPS).
Also ran a unigine benchmark and ended up at avg 78FPS with the standard settings how min FPS was only 19.2 (which seems low as other reviews report ~30FPS) …
Yes that is normal. AMD’s new PowerTune works by adapting core clock to meet Power and Thermal requirements. Using AMD overdrive you can increase your power and thermal limits to maximise the clock speeds. Providing you have the latest drivers (13.11 v 5) I can’t see you having any problems with performance but if you do you can normally fix these with OverDrive or another OC utility. Our temperature is a “worst case scenario” so 60 degrees under typical loads sounds about right. Remember this card is safe to operate up to about 80~90 degrees on the GPU core for extended periods of time.
Hi Ryan!
I wonder how much quieter the card gets by tuning the fan profile? Can you go below 50 db, and still have playable temperatures?
Sure you can. Our temperatures were the theoretical absolute maximum with default settings under furmark. Under typical gameplay 60~65 is more realistic (will vary by game). You could also tune fan profiles right down to very low levels to have like 75~85 under gaming and that would still be totally safe! I think 95 degrees is the TJ Max for the HD 7970 but im not 100%.
This fan with stock cooling is silent still at 45% fan speed. and that cools my (Overclocked to 1150gpu – 1600/6400 ram) R9 280x (7970 ghz edition) Vapor X just fine at 100% load.
in battlefield 4 i need to run on 100 godamn % what’s make alot of noise and every 2-3 hours required to cool or be shotdown i got a vaporX r9 280 psu 760w. what’s wrong ?
it’s get too hot
no shit
i have the same card and it get only to 64c on a 63% fan speed ….
change it in the store you purchase it or just use your warranty to change it
can some one please suggest if this card is good for 3d rendering on max with vrayrt… compared to the nvidia gtx series,
Nvidia is crap. they don’t have enough bus width and memory bandwidth so the bottleneck making the result lower than the AMD Radeon. Go for the AMD Radeon, you wont regret it. It’s the best gpu architecture in the world.
In my country, Sapphire’s Vapor X R9 280x OC is priced at around $394 whereas MSI GTX 770 is priced at $454. I’m leaning towards this card since it has 384 bit memory bus width and has 3GB VRAM. But some of the R9 280x owners are experiencing weird artifacts issues with their card while in-game as well as in 2D mode. suggest me which one to choose. If you say R9 280x, which brand or which model would you recommend?
I would suggest you to go for either MSI, Sapphire or XFX. so far these three brands doesn’t really experience nonsense artifacts. Those artifacts are just probably manufacturing defects. If budgets permits, go for the 290X or the 295X. Its worth the money especially for gaming and long term investment. Otherwise, stick with the 280X its more than enough for your everyday gaming.
I have use all top brands and models of 280x and found Powercolor Turbo Duo best.
Its best in performance, overclockability and rock solid looks and cooling.
Any card that shows defects is probably clocked beyond tolerances for that particular chip. This can be for a few reasons: 1. the person who owns it is overclocking the hell out of it causing it to artifact. 2. The manufacturer got overly ambitious when binning their chips and put lower performing chips in cards that were not meant for them just so they would have a high end product to sell, or a third possibility is that the owner is also using the card for mining and causing premature wear which is causing the artifacting when a game is fired up. This chip is the same chip that has been in use since the HD7950/7970 were introduced so AMD is getting good yields from them, I seriously doubt that there would be a major defect plaguing Tahiti chips this late in the game.
What about artifacting?
What are the r9 280x cards with the least artifacts appearing?
If the card is artifacting it is damaged and needs to be RMA. A normal 280x will not artifact.
“and it is on par with the GTX 770 which is about $100 more expensive.”
No, it is not on a par with the GTX 770. The GTX 770 only outperforms it in synthetic benchmarks. Nobody sits around running benchmarks all day. In games is where performance matters and in games R9 280x beats GTX 770, and in a few of your tests it even beats the GTX 780. R9 280x is far superior to GTX 770 when it matters, when playing games.
Not true, please educate yourself before commenting. Check out a wide variety of websites that will show you that the GTX 770 and R9 280X are on par – http://www.techpowerup.com/reviews/Sapphire/R9_290_Vapor-X/25.html Whether the R9 280X or GTX 770 wins will depend on: whether the game is Nvidia or AMD optimised, what the clock speeds of the respective cards are, whether both cards are using latest drivers.
I’m not referring to other people’s test results, I am referring to the test results that this site posted and then the erroneous conclusion that they came to when they have their own data that contradicts it. Other people’s test results are irrelevant. Their own charts show the R9 280x is superior to GTX 770.I am so sick of all the nVidia bias in the tech media. They run tests, AMD wins the tests, then they conclude that nVidia is best even though the numbers go against them. It’s been going on for years and it’s BS.
can i use enermax triathlor FC 550W on this card ?