|
Benchmark
Results :::...
TMPGEnc:::...

TMPGEnc
converts *.AVI file to MPEG1, the format which is used in
VideoCD. Using variety of option in TMPGEnc, you can compress
your video file in high quality.
TMPGEnc enables us adjust bitrate, quantize matrix, GOP
structure, Interlace and many other parameter so that you
can create most appropriate movie file depending on your
purpose.TMPGEnc, or "TSUNAMI-MPEG Encoder", as
it used to be called, TMPGEnc supports dual-CPUs for filtering
and part of the MPEG encoding procedure. 3D Now! and SSE
are also supported.
As
with every other time we've run this benchmark, a 692MB
*.avi file was encoded using the following parameters:

Needless
to say, the lower the encoding time the better the performance.
Here's how it all shped up:
TMPGEnc
- 32bit

Even
allowing for the fact that TMPGEnc takes advantage of the
X2's dual cores for filtering and part of the MPEG encoding
procedure, the FX57 shaves a significant percentage off
the overal encoding time.
TMPGEnc
- 64bit

Once
again we see the move from 32bit to 64bit leave the FX standing,
or more accurately going backwards. As both the X2 and 4000+
step up a gear, albeit not much of a gear in the X2's case,
the FX's rendering time increases from 414 seconds to 472
second.
PiFast:::...

- Computes
decimal digits of pi, e=exp(1), sqrt2, with the high limit
of 17 billion digits.
- Very
fast (fastest pi program available on the web - July 5
2003).
- Much
more economical in memory than other pi programs.
- Parameters
can be used to fit in the memory machine (memory saved
<-> time lost).
- Able
to handle disk memory for very large computations, permitting
to compute much more digits than could be contained in
the physical memory.
- Very
large computations can be done in several chunks.
- Can
compute a lot of other predefined constants (log(2), zeta(3),
...) with many formulas.
- User
can define his own formula to compute a constant.
- Can
compress directly the output result and later, decompress
all or part of it.
The
following method was used:

PiFast
- 32bit

PiFast
- 64bit

The
switch to 64bit does nothing for calculation speed, except
for the FX57 that is, which gets slower.
|