Can't help but notice that the GTX580 review got non of the naming scheme BS that the HD6870 got, for being a GTX485 and not a GTX580, and you even mention it's nothing new, but you focus (correctly) on the price not the BS names, just wish you could've seen fit not to waste a whole FAQin' page in the HD6870 review on that, but I guess the concern about people being fooled into thinking the HD6870 was a major upgrade for a much lower price was greater than the concern about the GTX580 being a major upgrade for the same price.
I guess when people come to expect bad naming for one IHV and not the other then it's worth a whole page to detract from the rest of the review and then in the other just ignore it as business as usual.
Wow you even managed to get in an Adobe CS5 plug with a hack for the one 1 out of 3 features MPE has that is CUDA accelerated (and to be replaced by OpenCL in the next release according to Adobe), all that before you were able to run more HQV or better tests on the HD6870 to finally improve on the rushed initial review that there was 'no time for'. And even in this test you half-ass it again !
"I’m leaving AMD’s cards out because this isn’t an AMD versus Nvidia battle. It’s all about hardware (GPU) against software (CPU) rendering—expect the AMD cards to look a lot like the Nvidia cards running in software mode."
If you're going to leave them out, then don't mention them, you don't even test the cards yet you comment on their expected performance in a way that is far from neutral, it clearly positions AMD v nV.
Jeez Chris, you should know better than that, seriously WTF kind of validity is there in that comment without actually making the test? Well we don't want to make it A vs B, but assume that B best effort is as bad as A horrible low-end results.
Cleeve should've included other references in his review for the HQV results and then use the time excuse when he could've simply added the provided material for reference in a shorter time than it took to write the naming BS (
no Cleeve telling me it gets 204/210 (or AMD got 198) doesn't mean Jack if I don't know what the intel or nV numbers are for the same rig/test, as they could be getting 207 and 209 respectively or did that not occur to you that without a relationship to other cards you don't know what a 'good', 'good enough' or 'great' score is).
Seriously guys, get some consistency in your testing methods and your writing. [:grahamlv:3]
And get a better editor who will force you guys to maintain internal and external validity in your testing methods, cause right now whether it's the excuse of time or effort, neither of these last two reviews have been up to the standards of even your Fermi paper-launch review and analysis in January.
For the limited time I have to read reviews, I see less reason to read a THG one as it's gone from a great technical review (Chris' old reviews were among the better ones out there) to simply a slight alteration on the re-hash of the PR kit. At least Crash/Thomas' reviews are still the A-class calibre he's been producing since before he joined the THG staff, I will still seek out his reviews as they remain among the best in the biz, even if he doesn't get all the attention and all the sexy new toys.
I'll wait for a better review (Like Xbit's that came a few days later) rather than worry about who published on day/minute one, and right now it seems that's the goal in order to get page views by getting it out there fast rather than getting out a review worthy of the previous THG efforts.
