Originally Posted by Alatar
Originally Posted by tpi2007
I'm still reading Anand's review but this slide does make things look rather disappointing in the CPU department (some previous paragraphs already hinted at it):
It's not even a straight 10%, it's ~10%, which probably means PR speak for anything between 9.5% and 9.9% rounded up.
It's a good performance improvement on its own, the problem is where AMD is coming from.
Problem being that frequency also went down 10% meaning that you get pretty much the same CPU power as with Richland.
For OCing it would seem like we need to wait for people to actually buy these since AMD didn't want to send out the samples early enough.
Haha, I just added that last part one or two minutes after writing it.
This leaves me wanting more. And continuously wondering what would have been if they had just concentrated on improving the CPU cores on Thuban / Llano instead of going this route. With each subsequent Bulldozer architecture improvement they are basically just playing catch-up with their own previous architecture.
Originally Posted by SKYMTL
Originally Posted by tpi2007
Except it won't be available at launch, which is quite weird.
There's a lot of things weird about this launch.
Like the broad-scale availability of the A10 7850K at retailers but AMD's flat out refusal to sample them to most sites for launch. Meanwhile, everyone got an A8 7600, an APU that won't be available for the next month or so.
Or the fact that we got the APUs TWO business days before launch, during CES.
This is what's called a targeted PR strategy and it happens all the time. How do you make sure a car reviewer doesn't find all the issues with a new model? Give him just 15 minutes of seat time and pop him into the trim level that's sure to meet a certain criteria.
That's exactly what happened here.
AMD sure knows how to be a bit weird. Let's see how this whole thing unfolds and if there is something going on with the A10-7850K's retail performance.
One of the main advantages of the Richland family compared to Trinity was that Richland CPU cores didn't downclock from their base clock
nearly as much as Trinity did under continuous heavy load. Let's hope they didn't do the same mistake again.
Originally Posted by sugarhell
Its kinda tricky to see the IPC increase because of the clocks. I am waiting for a better review/results from ocn to judge better
What's interesting is that they could have done one if they had an A10-6700 at hand (Edit: though the base clockspeed is the same, at 3.7 Ghz, the A10-6700 has a higher Turbo, at 4.3 Ghz, unless the Turbo speed is configurable, it wouldn't be a fair comparison either. In any case, downclocking an A10-6800K would achieve the same results).
In fact, along with the A10-6800K, the A10-6700, which is a 65w TDP part, is one of the most interesting parts from the Richland family, it's slightly faster than the A10-5800K in everything, on a 65w TDP.Edited by tpi2007 - 1/14/14 at 6:22am