New Posts  All Forums:

Posts by TranquilTempest

I think the two side boosters will usually land at the cape. The center core is more likely to either land on the ocean platform or be discarded, because you take a bigger payload hit to try boostback and recovery with it(especially if they decide crossfeed is worth doing).
Blower designs have been around a lot longer than vapor chambers. Vapor chambers are more expensive, so they're only used on high TDP parts. Hell, the reference 970 used heat pipes instead of a vapor chamber, and there's a good chance the 1070 will too.
20% is expected from hyperthreading on a 4 core processor when 8+ threads are heavily loaded and the game is still CPU limited. This is NOT a common scenario. Out of the thousands of games out there, how many can you list where this "20% to minimum framerates" applies? I bet it's less than 5. I said games could not easily spread their work out to many threads, I didn't say it was impossible. If it was easy, everyone would be doing it.
That would be down to more cache and higher stock clock speeds, not hyperthreading.
Hyperthreading only matters if you run out of real cores. Cinebench can easily spread its workload over 16+cores. Games can't. Hyperthreading rarely matters on 4 core intel chips, it will matter even less on an 8 core chip.
Maybe CB is just a favorable benchmark for Zen, maybe its stock clocks and TDP are so high it eats up all the overclocking headroom, maybe its not an accurate number, maybe it really performs that well, maybe it's priced like an 8 core Haswell part, maybe it redefines the CPU market.We can hope, but lets not get too carried away, AMD wants to make a profit here.
1kfps 1khz masterrace.
There are a whole lot of variables that determine whether you see a benefit. Most of this has to do with how accurate the CPU is at prefetching the information it needs, and whether the memory interface is saturated. If you're completely bandwidth limited, latency isn't going to matter much, if you have perfect prefetching, latency isn't going to matter much. If neither of these are true, latency matters just as much as bandwidth. If you're GPU limited, you should be...
TBH,There is no perfect place to get all your information, different reviewers test different things well. For example, techpowerup has the best methodology for testing power usage, but they don't do fcat or even frame times. The reviewers that do fcat don't test many titles, etc. Every reviewer that's the best at SOMETHING sucks at something else.
I believe old bridges will work at the old bandwidth.
New Posts  All Forums: