I think AMD's unlocked voltage control is better than Nvidia's boost from an overclockers point of view. However Nvidia's solution is much more precise and a far better solution when it comes to a hands off approach. AMD's isn't actually all that great, it's a lazy blanket solution that results in higher power consumption than should be taking place, as well as more degradation due to higher temps, more voltage applied of course.
However as someone who enjoys tinkering with hardware this wasn't a problem for me. Once I got a kill-a-watt I was able to see just how badly overvolted my card was at stock, being a boost TF3 I had stock clocks of 960MHz, with a stock voltage of 1.25v... Which needless to say unless you've got the crummiest gpu in the world, is well past what's actually needed to start with (later on you'll probably need to raise it, but out of the gate it's beyond overkill). Dropping my voltage at stock clocks from 1250mv to 1063mv, stock voltage actually used 22% more power in my tests than a manual undervolt... It's quite a considerable margin of error, consider that 1050/1500 clock speed with manual voltage resulted in less watts drawn than stock clocks at 1.25v you can see how this crude method really left a lot of performance on the table for AMD.
I don't have a Kepler card, and likely never will. I can't really comment on boost/boost 2.0 because of that, but the total lack of actual control presented by Nvidia was one of the major reasons I never purchased a Kepler card. I understand they want to cut into the console market, that's the direction they want to go... Which means software like GeForce Experience, and cards that auto overclock and are fool-proof for foolish overclockers. AMD probably understands the advantages this presents as well, which is why so many cards are now coming voltage locked out of the gate for AMD. I think between Kepler Boost and AMD locked voltage neither are desired, however in that situation I'd much rather have a Kepler because they're much more precise (seemingly) with their voltage/clocks regulatory system... It's probably not as accurate as manual control, but it has to be far and away better than AMD's solution to simply lock out any voltage control at all. It makes downclocking and undervolting a near pointless affair, and it really takes away the end users ability to hone their experience and be more cost effective with their hardware by delivering more structured clocks and voltage per need on a per game bases.
/Meh
Enthusiasts seem to be an after thought for both at this point, quite a disappointing generation tbh... If this were cpu's it'd be like Intel releasing 1155 without k chips at all for Nvidia, or in AMD's case, opening with all k processors and then slowly fusing them off as time goes on until you don't know which is which anymore.
However as someone who enjoys tinkering with hardware this wasn't a problem for me. Once I got a kill-a-watt I was able to see just how badly overvolted my card was at stock, being a boost TF3 I had stock clocks of 960MHz, with a stock voltage of 1.25v... Which needless to say unless you've got the crummiest gpu in the world, is well past what's actually needed to start with (later on you'll probably need to raise it, but out of the gate it's beyond overkill). Dropping my voltage at stock clocks from 1250mv to 1063mv, stock voltage actually used 22% more power in my tests than a manual undervolt... It's quite a considerable margin of error, consider that 1050/1500 clock speed with manual voltage resulted in less watts drawn than stock clocks at 1.25v you can see how this crude method really left a lot of performance on the table for AMD.
I don't have a Kepler card, and likely never will. I can't really comment on boost/boost 2.0 because of that, but the total lack of actual control presented by Nvidia was one of the major reasons I never purchased a Kepler card. I understand they want to cut into the console market, that's the direction they want to go... Which means software like GeForce Experience, and cards that auto overclock and are fool-proof for foolish overclockers. AMD probably understands the advantages this presents as well, which is why so many cards are now coming voltage locked out of the gate for AMD. I think between Kepler Boost and AMD locked voltage neither are desired, however in that situation I'd much rather have a Kepler because they're much more precise (seemingly) with their voltage/clocks regulatory system... It's probably not as accurate as manual control, but it has to be far and away better than AMD's solution to simply lock out any voltage control at all. It makes downclocking and undervolting a near pointless affair, and it really takes away the end users ability to hone their experience and be more cost effective with their hardware by delivering more structured clocks and voltage per need on a per game bases.
/Meh
Enthusiasts seem to be an after thought for both at this point, quite a disappointing generation tbh... If this were cpu's it'd be like Intel releasing 1155 without k chips at all for Nvidia, or in AMD's case, opening with all k processors and then slowly fusing them off as time goes on until you don't know which is which anymore.
Facebook
Twitter