That is excellent.  Funny, I was just reading Toms Hardware's latest 
test of the 939 pin Athlon64 thinking that none of these benchmarks 
really relate stuff I care about and how to get benchmarks of P&R. 
Michael, thank you soo much!

Indeed, Quartus II is really the only app I run that takes enough time 
to be annoying; even my toy designs take more like 5 minutes to 
complete. (Games aren't even that CPU limited anymore).

Ace's result are a bit surprising.  Ignoring Intel, more memory 
bandwidth makes nearly no difference and twice the L2 doesn't make up 
for 9% frequency loss, so it looks like frequency is the factor to 
optimize for this application (ie., the 3700+ beats the 3800+).

Oh, and no surprise, AMD wins on price/preformance big time.

Tommy


Michael S wrote:
> Computer hardware enthusiast Johan De Gelas of aceshardware compared
> performance of seven top PC MPUs on Quartus-II Web Edition place and
> route task.
> Johan is a real professional when it comes to benchmarking computers.
> His results can be safely relied on.
> 
> Probably the article would be of interest for participants of
> comp.arch.fpga
> 
> http://www.aceshardware.com/read.jsp?id=65000307
> 
> Johan recognizes that the test design was uncharacteristically small
> by modern FPGA's standards. He is willing to improve the benchmark.
> Please, post you proposals for improvements in the aceshardware
> message board.
> 
> http://www.aceshardware.com/forum
> 
> Just keep in mind that 
> A. Benchmarks have to utilize free or donated tools and free or
> donated design files.
> B. Johan is not logic designer. Try to make benchmarks setup easier
> for him. He has more than enough head ashes with the hardware he is
> testing, doesn't have to mess with software as well.
> 
> Regards,
> Michael
> 
> P.S.
> Many thank to Zhao Ming, who supplied radix-4 CFFT core that served as
> the basis for the benchmark.
> And of course to Altera Corporation for free Quartus tools.

Computer hardware enthusiast Johan De Gelas of aceshardware compared
performance of seven top PC MPUs on Quartus-II Web Edition place and
route task.
Johan is a real professional when it comes to benchmarking computers.
His results can be safely relied on.

Probably the article would be of interest for participants of
comp.arch.fpga

http://www.aceshardware.com/read.jsp?id=65000307

Johan recognizes that the test design was uncharacteristically small
by modern FPGA's standards. He is willing to improve the benchmark.
Please, post you proposals for improvements in the aceshardware
message board.

http://www.aceshardware.com/forum

Just keep in mind that 
A. Benchmarks have to utilize free or donated tools and free or
donated design files.
B. Johan is not logic designer. Try to make benchmarks setup easier
for him. He has more than enough head ashes with the hardware he is
testing, doesn't have to mess with software as well.

Regards,
Michael

P.S.
Many thank to Zhao Ming, who supplied radix-4 CFFT core that served as
the basis for the benchmark.
And of course to Altera Corporation for free Quartus tools.