ASUS R9 290X DirectCU II OC Review

[Master&Puppet]

I think ASUS have done a good job here but no less than I would have expected from any of the GPU partners. I think it is over priced though £499 RRP is very near to the bottom of the 780ti territory and I'd choose one of those for an extra few quid over this. It's too large a mark up for a cooler for me.

You can get a full EK water block cover for a 290x for virtually the same price as this and that is leagues ahead again if you've already got a custom kit for your CPU like many of us do. £440-470 is a more sensible price range IMO although getting free games does help quite a bit.

I don't see why people are getting upset about the 'heat issue'. AMD designed these cards to run hot. That means that it is warrantied to run at 95C at full load for the length of the warranty period or they'll have to give you a new one. Likely scenario is that it will outlive its usefulness before it breaks.

As for worrying about dumping hot air into the case for the rest of the components to absorb? Well its TDP is only 40W greater than a 780 which can't account for more than a couple of degrees and that will only make a difference if all your components are already running on their limits too.

 

[kinggavin]

crysis 3 2560x1440 very high with 8xmsaa , 30 fps is console fps, £500 price tag and newest card you would expect 60fps

 

[SeekaX]

crysis 3 2560x1440 very high with 8xmsaa , 30 fps is console fps, £500 price tag and newest card you would expect 60fps

this barely even deserves a reply.
a console doesn't run the game at 1440p, it runs the game at 720p, probably even lower. it also doesn't use any AA or anisotropic filtering.

 

[Ghosthud1]

crysis 3 2560x1440 very high with 8xmsaa , 30 fps is console fps, £500 price tag and newest card you would expect 60fps

consoles also use 1024x1024 textures, lower resolutions and considerably lower hardware level access. You cant even begin to compare them on the same level crysis is known to use 2K and 4K textures and is scalable to the engines maximum outputs such as ultra details. Consoles use custom settings somewhere in the range of low/medium settings.

 

[Mishan]

Not sure I trust the benchmarks of this site.

Look at Crysis 3.

The overclocked(!) card is doing a LOT worse than the same card with stock speeds on 1440p. We're talking 25% or so difference.

Also, just focusing on FPS is misleading, FCAT matters as much if not more.
Finally, PCPER.com showed that due to much improved thermals, there was no throttling, giving you a 17% boost in clock speeds. This made the card beating or running even with the 780 Ti(stock) in 7 out of 8 games at 1440p.

Yet these guys made no real research on the throttling issue, no FCAT, frame rate stability and have really weird benchmarks(like the Crysis 3 one where the stock clocked card does much better than the same card with OC on 1440p). Did anyone not catch this before publication?


Embarrassingly bad "review" full of holes.

 

[SieB]

Looks like the Crysis 3 OC and stock results for 1440p are the wrong way around but the 8fps gain mentioned for the OC in 1080p isn't shown in the graph though, both stock and OC show 54fps.

 

[barnsley]

Also, just focusing on FPS is misleading, FCAT matters as much if not more.



Embarrassingly bad "review" full of holes.

FCAT, the Frame Capture Analysis Tool, is NVIDIA’s take on what the evolution of frame interval benchmarking should look like.

Because using a tool made by Nvidia won't be biast at all...

 

[NeverBackDown]

consoles also use 1024x1024 textures, lower resolutions and considerably lower hardware level access. You cant even begin to compare them on the same level crysis is known to use 2K and 4K textures and is scalable to the engines maximum outputs such as ultra details. Consoles use custom settings somewhere in the range of low/medium settings.

First off consoles have the same hardware level access and if not much higher access. Consoles are set hardware vs PCs ranging from billions of different configurations thus allowing the devs to push the console hardware efficiently vs PCs getting less efficiently and making the drivers do most of the work(which is bad because they are inaccurate).

Consoles use a mix of settings and depends on which console to be exact. PS4 runs the same settings but runs at a higher resolution in the case of BF4 but in the case of AC4 it uses same 1080p res but higher quality settings.


Anyway back on topic.... DCU2 is a great card but i still think no 3rd party air cooler is enough for the 290x. It would need a couple 10mm copper heatpipes to be able to keep temps very low.

 

[Master&Puppet]

crysis 3 2560x1440 very high with 8xmsaa , 30 fps is console fps, £500 price tag and newest card you would expect 60fps

It's not as simple as that dude.
Of course consoles won't run the games at any where near that resolution. Look at BF4 on the PS4 - they had to choose between 1080p 30fps or upscaled 720p at 60fps (they wen't with the latter), that's pretty lame imo. And if you think this is bad then don't check out the 780ti which costs another £100 for another 5fps lol.

As for:

Not sure I trust the benchmarks of this site.

Look at Crysis 3.

The overclocked(!) card is doing a LOT worse than the same card with stock speeds on 1440p. We're talking 25% or so difference.

Also, just focusing on FPS is misleading, FCAT matters as much if not more.
Finally, PCPER.com showed that due to much improved thermals, there was no throttling, giving you a 17% boost in clock speeds. This made the card beating or running even with the 780 Ti(stock) in 7 out of 8 games at 1440p.

Yet these guys made no real research on the throttling issue, no FCAT, frame rate stability and have really weird benchmarks(like the Crysis 3 one where the stock clocked card does much better than the same card with OC on 1440p). Did anyone not catch this before publication?


Embarrassingly bad "review" full of holes.

You haven't been around here long enough to make sweeping judgments and attacks like this...

Firstly. Assuming that the Crysis 3 figures aren't simply the wrong way around... It is perfectly possible for an overclocked card to perform worse in a game than at stock. You do see it quite often when reviews benchmark a card across a large number of games because the team will push the card as far as they can and it will be stable in almost all situations but not necessarily in all. It can be than in Crysis the card was pushed a little too far and the FPS suffered. Furthermore in this situation it probable that overclocking the card just made it run hotter which simply caused more throttling and therefore actually scoring lower FPS. There's a lot to consider... than just "UH OVERCLOCK HAS TO WIN!!11".

Secondly FCAT is interesting and all but has almost no significance to single card configurations. Ever since Nvidia started pushing the FCAT software it was obvious that AMD was only badly affected in multi-card configurations and this has largely been solved in the recent AMD driver roll outs (although work needs to be done in eyefinity xf conditions) so it is pretty easy to see why a smaller reviewing site such as OC3D wouldn't bother with such a large investment in FCAT software and hardware when the truth is that the AMD frame rate issue just isn't important enough these days to warrant exploring every game in every review with FCAT analysis. FCAT is largely a waste of time and money for smaller sites.

Thirdly if you actually bothered to watch the review for the r9 290x reference rather than read the summary (which is all the written part is) then all you'll hear it Tom complaining about the cooler lol...