06-07-2016 09:48 PM - last edited on 03-06-2024 01:48 AM by ROGBot
06-07-2016 11:27 PM
Gobe wrote:
Well, that sucks. I've been far too busy the past 6 weeks to look at Broadwell-E or play with benching, so I guess I'll spectate while this hopefully gets worked out.
06-08-2016 08:44 AM
Menthol wrote:
I suppose we'll see an update soon, can you run the stability test or since it doesn't recognize the chip it won't run at all?
06-09-2016 08:06 AM
Menthol wrote:
Good that means my score will remain at top a little longer:)
06-14-2016 05:58 PM
cekim wrote:
If programmers haven't gotten the message that GHz are dead and they need to learn how to parallelize in ways they've previously assumed impossible, then need to listen harder.
06-15-2016 08:55 AM
Gobe wrote:
Individually, silicon transistors with far faster switching speeds can be produced, but try to put millions of these on a single wafer and it's not so easy. GHz isn't dead but the headroom is running out under current technology. It's going to take some hard work to push silicon further.
Semiconductors with far better electron mobility might be the place to pick up (Gallium Nitride comes to mind). Rather than continuing to flog silicon to within an inch of it's life, it makes sense to work with a material where electrons can be moved more quickly with far less effort (read voltage and ultimately power) required to move them. Faster switching/lower power. Yes please!