A much better approach than the pseudo-advertising that HH tends to indulge in - add an AMD Gaming Evolved title with GCN super-friendly compute settings for AMD cards, 3D benching for Nvidia cards for example, and Batman AC shown as a comparison even though the game i.q. levels aren't equal*, and their perceived value based on their pricing ( the Devil 13 rated 8/10 even though no price is listed) that seldom bears any relation to reality.
10th May Nvidia card - Batman AC included
17th May AMD card- Batman AC not included
18th May Nvidia card - Batman AC included
22nd June AMD card- Batman AC not included
HI all, thought I would drop in and address a few comments from this thread to give a better understanding of our review (HH).
Regarding game choice, we always tend to go with a selection of the latest games where possible although another consideration is the genre/engine. For example I would always want to have a couple of FPS complimented by a racing, RTS, MMO etc rather than FPS after FPS.
Looking at Batman specifically (the original and the sequel), they were included in a batch of reviews when new but then at later dates moved to be PhysX only content which is why they were missing from some AMD reviews quoted above. Had a newer game happened to support PhysX it would have A) met our criteria for testing the latest games and B) been the PhysX test at the same time. We wouldnt normally do a NV review (or AMD) with a key feature test missing...reviews are about educating people new to the product as well as showing performance. (Borderlands 2 will be added next month, and will drop out other than some specific tests when something similar/newer comes along)
We also don't pick games to favour one manufacturer or the other. Dirt Showdown, as one example, happens to perform best on AMD but it is also the newest "big" racing title around and benefits from being DX11 etc which is also taken into consideration. F1 2012 will replace it, unless something better suited comes along in the racing genre. We used to use NFS but recent games havent had the best engine for real world benchmarking.
That brings me on to the next point, the games themselves play a big part. This review was the first to use Sleeping Dogs. I also took a look at Dark Souls for possible inclusion but removed it due to it being a really dodgy console port (engine wise, the game is great on 360). Frame caps at 30fps being a big problem. Skyrim nearly made it in but for something going wrong with the game install.
It's probably also worth noting that I agree 100% with W1zzard on comparing reviews. It just isn't possible to do that. We for example retest every result in each review on the days running up to the review with the latest drivers, patches etc for all cards included. This review it was only 2 cards (excluding 3DMark) but it can be several. For this review we also OC'd our CPU, to 5GHz, only used the OC Devil BIOS, use a higher spec CPU, more memory, different chipset architecture, different game segments the list goes on when compared to W1zzards work.
There is no right or wrong some will prefer our way, some TPU's. I would say its best to read both (and more) then form a decision based on all of them combined.
I hope that gives a better understanding about how we choose the games included, and the process as a whole.
Finally, its worth noting that the card we received had no heatsink/thermal issues. It worked properly right out of the box.
Of course CF support is a mess most of the time, you're better off watercooling a single card and overclocking it to 1300 MHz. Haven't really tried SLI so can't comment on that side.
