Far Cry 2 Hardware Performance Review Comment Thread

[FiXT]

'Better late than never" they say.

Unfortunately we were a bit slow on the uptake, but hopefully we have been able to make up for it with the thoroughness of the review. Having taken many of your comments into consideration we have attempted to adapt Hardware Performance Reviews into something that is chalk full of information and entertaining.

As you will note, the format has been altered slightly to eliminate tons of scrolling - hopefully this is a change for the positive.


As usual - feedback is more than appreciated. Questions, comments, recipes - feel free to bring them forward.

Hope you enjoyed - And Don't forget to play on Far Cry 2 in DirectX10!

You can check out the review here: Far Cry 2 Hardware Performance Review

[Chilly]

Great review, really informative, LOVE the tabbed feature/way of showing the information. Only thing I disliked was the different scales on the result pages when comparing DX10 - DX9. For example on the 1920x1200+Ultra High Graph, the DX10 graph scales from 0->70, but then the DX9 graph scales from 0->80. This gives it the visual impression that DX9 is MUCH slower than it really is.

Please keep the scales the same if done in this "tab" format, otherwise it throws people off, and feels a bit deceptive. That's my ONLY complaint thou, great content, lots of great info! Good job guys!

[Eagle Eye]

I agree with Chilly love the tab features, but honestly I think i would have to see it to believe that there is some notable difference between DX9 and DX10. Its also nice to see that the new drivers from nVidia actually help, not that performance was suffering substantially from my current rig to even my AMD rig which ran most games quite well.

EE

[gs01]

The frames might be higher in dx10 but the minimum framerates in dx10 are much worse then dx9 and the minimum is what matters the most, so I say dx10 still fails.

[tom.mullie]

just a comment about the CPU results,
You say that multi-core processors show real benefits. correct me if I'm wrong, but the CPUs actually finished more or less in order of clock speed, didn't they. The e8400 came second, and not by a whole lot (considering being down two cores)

Also, the tab thing is great. way better than the old way, especially for comparing screenshots

[Squeetard]

How about a link to the article?

[Squeetard]

Originally Posted by tom.mullie

just a comment about the CPU results,
You say that multi-core processors show real benefits. correct me if I'm wrong, but the CPUs actually finished more or less in order of clock speed, didn't they. The e8400 came second, and not by a whole lot (considering being down two cores)

Here's how I read it. The identically clocked quad core is 21.5% faster than the dual core. To me that's a big difference.

[FiXT]

Top of the morning!

Good thinking Chilly - While the differently scaled graphs work when comparing the chart numbers to itself, for comparing DX9 and DX10 it would may make things much clear keeping it the same - Will play around with it a bit and hopefully implement that soon.

As one set of the screenshots show, there really is very little, if any difference between DX9 and 10, so either way your going to get virtually the same quality, however, it may mean the difference between playing at 30FPS+ or having to deal with choppiness - so it depends on the situation. For those running Vista, the pros far outweigh the cons of DX10.

@tom.mullie

You are right, that the CPU's are - for the most part - in order of clock speed, however you must isolate the results.
The Q9650 which is a 45nm quad core clocked at 3.0GHZ hit 48FPS, the E8400 which is a 45nm dual core clocked at 3.0GHz hit 39.5FPS which means the identically clocked quad core netted a 20% performance boost.

Also in the AMD realm we see the Tri-Core X3 8750 (2.4GHz) and Quad Core X4 9850 (2.5GHz) hitting equal and higher numbers than the 3.2GHz X2 6400+, now while some of this may be due to the improved Phenom architecture, I also stems from the multi core element.

Hope this info helps :)

[MpG]

Nicely done again. You obviously put some finsane work into these reviews, but the it's hard to argue with the final result!

About the only real things I can complain about is that I'm a big fan of minimum frame rate reporting, since there are often significant differences to be found at times. Also, the height of the graphs - while being able to switch from DX9 to DX10 with the bottom tabs is handy, the height of the graphs mean I have to scroll up and down to compare the higher end cards. Maybe split the graphs into two chunks? But yeah, if that's the best I can do for complaints...

Any ideas on ATI's multi-GPU scaling? It seems to be all over the place. Sometimes you get near-zero improvement going from a 4870 to an 4870X2, yet great scaling after you add another X2? Other times, there's great improvement from the 4870 to the X2? Weird. Common wisdom suggests that it should be a matter of diminished returns as you add more and more cores, but the graphs suggest something else is going on?

[FiXT]

I like minimum frames and actually have all the numbers, however - as you mentioned with the scrolling problem, the graphs are already 1055 pixels high, adding in minimum frames put them well over 1500 which means a LOT of extra scrolling - so for simplicity sake it was left out.

In regards to the ATI situation, from what I could tell in doing the tests there is a combination of factors. One of which is drivers. The ATI drivers were buggy, in fact we had to omit Crossfire results for the 4870 and 4850 cards because for one reason or another it performed absolutely flaky- the 4870X2 on the other hand performed just fine in crossfire?! As a reviewer you learn to occasionally avoid asking why and wasting time trying to fix any one of hundreds of problems and just accept the fact that things are just not working :P
The other is that DX10 appears to be more CPU bound, as such you get more bottlenecking when running in DX10, thereby diminishing GPU frames; whereas DX9 is seemingly a little easier on the system and as such allows the GPU's (at least with ATI and their drivers) to perform better.