Review Request: Backport some Hotspot defaults for AMD Bulldozer processor

[Vladimir Kozlov]

Changes look good, I verified that they match the combination of next changes in 
jdk7:

7037812: few more defaults changes for new AMD processors
7035713: 3DNow Prefetch Instruction Support
7032133: Enable sse4.2 for new AMD processors

I filed bug:

7108221: Backport to jdk6 Hotspot defaults for AMD Bulldozer processor

I am running changes through our testing system. After it finished I will push 
it into http://hg.openjdk.java.net/jdk6/jdk6/hotspot.

Regards,
Vladimir

Deneau, Tom wrote:
> Please review this patch which backports some Hotspot defaults for AMD's 
> new Bulldozer which had already been accepted into Hotspot 21 in JDK7 
> General Release.  These defaults allow the Bulldozer processor family to 
> get the best performance.
> 
> The webrev is at
> 
>   http://cr.openjdk.java.net/~tdeneau/openjdk6-bd-defaults/webrev.01/ 
> 
> The changes consist of:
> 
> ·         default for AllocatePrefetchStyle goes to 0, that is, no SW 
> Prefetch
> 
> ·         UseXMMForArrayCopy, UseUnalignedLoadStores and 
> UseSSE42Intrinsics default to true
> 
> ·         changes the logic for detecting whether or not a processor 
> supports the 3dnow prefetchw and prefetch instructions.   A separate 
> CPUID bit (defined in about 2007) allows a processor to support 3dnow 
> prefetch instructions without supporting the whole 3dnow instruction 
> set.  The Bulldozer processors from AMD are the first that support 3dnow 
> prefetch without supporting the 3dnow instruction set.  The logic change 
> is really one small change in src/cpu/x86/vm/vm_version_x86.hpp but to 
> clarify things I changed a function name from supports_3dnow() to 
> supports_3dnow_prefetch() which is really what was meant all 
> along.  This was the reason the other files changed.  I did not make any 
> change in src/cpu/x86/vm/x86_64.ad since that was not checking for 3dnow 
> support.
> 
> -- Tom Deneau, AMD
> 
>  
>