So ROLLO, how do you feel about the 5800 ultra automatically throttling it's clock speeds down from 500/500 to 300/300 when it get's hot (re: when you are gaming more than 5 minutes)? I've just read yet ANOTHER review today where they mention that the card automatically throttled the clock speeds down while they were in the middle of running a benchmark. Actually the one today said it happened multiple times. They said the benchmark numbers dropped 30% when this happened. Did they include those benchmarks in the review though? NO. Just like every other site I've seen thats said the card has automatically underclocked itself while running benches, they threw the score out and reran the bench until it made it through without throttling itself and used that score. So Rollo, those benchmark scores you are seeing, don't count them for a representation of what actual gaming will be like. You can bet if a 3 minute benchmark overheats the card enough to make it throttle itself down, an hour of gaming will certainly cause this to happen multiple times. Can you imagine the dissapointent you will get when you spend $400 for this card because of the wonderful bench scores it got in the reviews, yet find out it keeps clocking down to 300/300 when you actually play a game for more than 5 minutes?
Looking at the 9700 and the nv30 the 9700 running at 325/310 is running neck in neck with the nv30 running at 500/500. What do you think happens when the nv30 clocks itself down to 300/300? Thats 25mhz SLOWER than the 9700 core and 10 mhz SLOWER than the 9700 memory. If the Nv30 runs neck in neck with the 9700 with a 175mhz advantage it's getting eaten for lunch when it throttles itself down to being slower than the 9700.
I've got a BIG problem with this. First off, I don't think the reviewers should keep reruning the bench till they get the score they want. If the nv30 downclocks and has a crappy bench score, it should be used. Why? Because that will be indicitave of what actual gameplay will be like. Don't you remember Nvidia making a big stink about 3dmark03 because it wasn't indicitave of actual gameplay? Well neither are these fraudulent bench scores that had to be reached by continually having to let the card cool down and rebench.
The second problem I have with this is that NOBODY is making much of a big deal about this. I've seen this Casually mention in MULTIPLE reviews now, that they had problems with the card automatically slowing down. This is not something that should casually be mentioned. This is something that needs to be stressed. Who the heck wants to deal with a card that automatically slows down when the gaming is getting intense and the card gets hot?
Another problem I have is with all of the 5800 ultra reviews period. Why are all the sites comparing an ultra card that will only be available in VERY LIMITED supply. (only those who've preordered will ever see one) to the 9700 pro. Why have seen NOT ONE 5800 non ultra review? Why has Nvidia not made any of these available for review? The only card made available for review is a card that very few people will actually be able to buy.
So now Huang and the Nvidia PR machine (oh, and you too rollo) are running around claiming they hold the performance crown. What a joke. They claim the crown with a card that has to be rebenched multiple times to get the desired performance on a card that will not be available.
Enjoy your fx Rollo. Hope you don't mind paying $400 for a 500/500 card that will keep dropping to 300/300 when you are in the heat of battle. That's a 40% reduction. Ridiculous.
Looking at the 9700 and the nv30 the 9700 running at 325/310 is running neck in neck with the nv30 running at 500/500. What do you think happens when the nv30 clocks itself down to 300/300? Thats 25mhz SLOWER than the 9700 core and 10 mhz SLOWER than the 9700 memory. If the Nv30 runs neck in neck with the 9700 with a 175mhz advantage it's getting eaten for lunch when it throttles itself down to being slower than the 9700.
I've got a BIG problem with this. First off, I don't think the reviewers should keep reruning the bench till they get the score they want. If the nv30 downclocks and has a crappy bench score, it should be used. Why? Because that will be indicitave of what actual gameplay will be like. Don't you remember Nvidia making a big stink about 3dmark03 because it wasn't indicitave of actual gameplay? Well neither are these fraudulent bench scores that had to be reached by continually having to let the card cool down and rebench.
The second problem I have with this is that NOBODY is making much of a big deal about this. I've seen this Casually mention in MULTIPLE reviews now, that they had problems with the card automatically slowing down. This is not something that should casually be mentioned. This is something that needs to be stressed. Who the heck wants to deal with a card that automatically slows down when the gaming is getting intense and the card gets hot?
Another problem I have is with all of the 5800 ultra reviews period. Why are all the sites comparing an ultra card that will only be available in VERY LIMITED supply. (only those who've preordered will ever see one) to the 9700 pro. Why have seen NOT ONE 5800 non ultra review? Why has Nvidia not made any of these available for review? The only card made available for review is a card that very few people will actually be able to buy.
So now Huang and the Nvidia PR machine (oh, and you too rollo) are running around claiming they hold the performance crown. What a joke. They claim the crown with a card that has to be rebenched multiple times to get the desired performance on a card that will not be available.
Enjoy your fx Rollo. Hope you don't mind paying $400 for a 500/500 card that will keep dropping to 300/300 when you are in the heat of battle. That's a 40% reduction. Ridiculous.