J
johns
Previous post on Dell 4600 giving an image data rate
of 1.4 gig/min ( very very high ) ...... Talked to another
guru at Intel. He says "no way hyper-threading is
responsible for this benchmark" DR DOS is not hyperT
aware. It has to be the ddr2 ram. His explanation I did
not completely understand. Basically he said that even
though any ram is much faster than the image benchmark
of 1.4 gig/min, there is the problem of "latency" along
the way with other pieces of hardware. The big war going
on now in all the recent developments .. hyper-thread,
hyper-transport .... is to solve the external hardware
latency that bottlenecks data transfer ... esp large amounts
of data going to video cards and hard drives. You can
buy a hot-dog system, that will stall bigtime in certain
apps, and run like mad in others. I use the game Far
Cry as a benchmark for any new systems that come my
way. My AMD64s / ATI 9800 Pro , run FC very well.
However, I have discovered that certain services in
those systems will cause stuttering and limited AI in that
game. Those are "services" installed by web sales types
pushing pop-ups, etc, and I eleminate them to make
the game run in top form. On the Dells, Far Cry will not
even start. The integrated video is not enough. I have a
Dell 4700 / nvidia 6800 coming in. I'll benchmark that
to see if indeed the AMD64 is still king of the hill in
overall performance. The AMD64s answer to hyper-
threading will be dual core processors. For those of
you who want to get past the stupid know-it-alls, I
have two pretty good "intuitive" benchmarks ...
PowerQuest Disk Image 2002, and the first Far Cry
demo. DI will give an honest appraisal of "latency"
in your hardware, and Far Cry will certainly test your
video quality and drivers. I have other benchmark
programs, but frankly I find them to be nerdy and
give little, if any, useful information ... with the possible
exception of 3DMark2002, but that is just a video test
of raw speed and badly dated. The new 3DMks are
meaningless unless you have a bunch of systems to
bench against. I have 1200. The problem of latency
as of right now, has not been solved very well. You
can pay a lot of money for nothing based on benchmarks
and hype from the know-it-alls. There are plenty of
fast components, but they don't link up well .. yet.
johns
of 1.4 gig/min ( very very high ) ...... Talked to another
guru at Intel. He says "no way hyper-threading is
responsible for this benchmark" DR DOS is not hyperT
aware. It has to be the ddr2 ram. His explanation I did
not completely understand. Basically he said that even
though any ram is much faster than the image benchmark
of 1.4 gig/min, there is the problem of "latency" along
the way with other pieces of hardware. The big war going
on now in all the recent developments .. hyper-thread,
hyper-transport .... is to solve the external hardware
latency that bottlenecks data transfer ... esp large amounts
of data going to video cards and hard drives. You can
buy a hot-dog system, that will stall bigtime in certain
apps, and run like mad in others. I use the game Far
Cry as a benchmark for any new systems that come my
way. My AMD64s / ATI 9800 Pro , run FC very well.
However, I have discovered that certain services in
those systems will cause stuttering and limited AI in that
game. Those are "services" installed by web sales types
pushing pop-ups, etc, and I eleminate them to make
the game run in top form. On the Dells, Far Cry will not
even start. The integrated video is not enough. I have a
Dell 4700 / nvidia 6800 coming in. I'll benchmark that
to see if indeed the AMD64 is still king of the hill in
overall performance. The AMD64s answer to hyper-
threading will be dual core processors. For those of
you who want to get past the stupid know-it-alls, I
have two pretty good "intuitive" benchmarks ...
PowerQuest Disk Image 2002, and the first Far Cry
demo. DI will give an honest appraisal of "latency"
in your hardware, and Far Cry will certainly test your
video quality and drivers. I have other benchmark
programs, but frankly I find them to be nerdy and
give little, if any, useful information ... with the possible
exception of 3DMark2002, but that is just a video test
of raw speed and badly dated. The new 3DMks are
meaningless unless you have a bunch of systems to
bench against. I have 1200. The problem of latency
as of right now, has not been solved very well. You
can pay a lot of money for nothing based on benchmarks
and hype from the know-it-alls. There are plenty of
fast components, but they don't link up well .. yet.
johns