CASTalk.com

Chris Thomasson · Guest

[quote]average speedup: 66,09%

An average speedup of 66% - that's really amazing!
[/quote]
Humm... Yes, the numbers do look promising...

Oliver S. · Guest

[quote]As I have some doubts that Chris is able to apply the fix
mentioned in my recent posting:

Ummm, don't worry Oliver, I can understand your code!
[/quote]
Ok, then I'm wrong here;
but that's just because your code looks like you've a knot in your brain.

[quote]Humm... Did you get beat up a lot as a kid?
[/quote]
Every day!

[quote]You seem to need, an attitude adjustment.
Now I think I know why you frequently use X-No-Archive: Yes.
[/quote]
Yes, _you_think_ that you know it!

Oliver S. · Guest

[quote]Humm... Yes, the numbers do look promising...
[/quote]
I'll do some other benchmarks, f.e. one with one thread that does SSE-operations
with some loads and stores and another thread that does integer-operations so that
the threas would mainly hit different execution-units. Wouldn't wonder me if this
woould give similar results; and that would indicate that when you run two diffe-
rent workloads, i.e. a rendering-application and a data-compressing tool, you'd
gain a higher throughput-improvement than with symmetric benchmarks. Wouldn't you
do some investigations or even build such benchmarks?

Chris Thomasson · Guest

[quote]Ummm, don't worry Oliver, I can understand your code!

Of course you can! That's just because it's modelled so exemplary.
[/quote]
:)

hyperthreading gains with overlapping memory-loads Goto page Previous 1, 2

	CASTalk.com Forum Index -> Computer Architecture	All times are GMT Goto page Previous 1, 2
Page 2 of 2