You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
R
Robert Miles
parallax-scroll said:
That website doesn't make it clear whether Intel's efforts will slow
down medical research by decreasing the percentage of all video
boards which are Nvidia boards now usable to support the few
BOINC projects which are able to use video boards to do some
types of computer work faster than CPUs can do it, or whether
Intel plans to eventually speed up medical research by offering
software that makes it easier to write programs that run on
video boards using Intel's video chips but are doing things unrelated
to producing graphics.
http://www.gpugrid.net/
Robert Miles
F
Folk
That website doesn't make it clear whether Intel's efforts will slow
down medical research by decreasing the percentage of all video
boards which are Nvidia boards now usable to support the few
BOINC projects which are able to use video boards to do some
types of computer work faster than CPUs can do it, or whether
Intel plans to eventually speed up medical research by offering
software that makes it easier to write programs that run on
video boards using Intel's video chips but are doing things unrelated
to producing graphics.
One sentence. Bet you can't say it without breathing.
R
Robert Myers
That website doesn't make it clear whether Intel's efforts will slow
down medical research by decreasing the percentage of all video
boards which are Nvidia boards now usable to support the few
BOINC projects which are able to use video boards to do some
types of computer work faster than CPUs can do it, or whether
Intel plans to eventually speed up medical research by offering
software that makes it easier to write programs that run on
video boards using Intel's video chips but are doing things unrelated
to producing graphics.
http://www.gpugrid.net/
It's a safe bet that Intel, just like IBM, is interested in any high-
profile application that will get them more visibility. This chip is
aimed at many markets, including HPC and other applications in science
and engineering. Graphics just happens to be the easiest to make a
business case for.
Robert.
Y
Yousuf Khan
Robert said:That website doesn't make it clear whether Intel's efforts will slow
down medical research by decreasing the percentage of all video
boards which are Nvidia boards now usable to support the few
BOINC projects which are able to use video boards to do some
types of computer work faster than CPUs can do it, or whether
Intel plans to eventually speed up medical research by offering
software that makes it easier to write programs that run on
video boards using Intel's video chips but are doing things unrelated
to producing graphics.
http://www.gpugrid.net/
Robert Miles
There's another development happening, based around the OpenCL
standards. AMD just recently showed gaming physics demo based around
both Havoc and OpenCL. Though this is based around gaming applications,
OpenCL is more general purpose than that.
Yousuf Khan
ATI ends the physics argument - The Inquirer
"Don't underestimate how big a deal this is, however. As soon as it is
optimised correctly, you can parse the physics load between the CPU and
GPU. If you have more of one than the other, you can still use physics
in the way it was meant to be played. Oops, wrong slogan... but this
implementation should actually do what the other side promises. The
upshot is that game developers can use physics more liberally, they
don't have to worry about minimum specs as much."
http://www.theinquirer.net/inquirer/news/673/1051673/ati-physics-argument
R
Robert Myers
There's another development happening, based around the OpenCL
standards. AMD just recently showed gaming physics demo based around
both Havoc and OpenCL. Though this is based around gaming applications,
OpenCL is more general purpose than that.
A better link to OpenCL without the Inquirer's usual slant:
http://en.wikipedia.org/wiki/OpenCL
People who write to specific platforms will always have to count on
the continued availability and competitiveness of the platform. x86
is number one in both of those categories, and it's the only plausible
candidate to ride out the microprocessor revolution. I'd bet on it
again at this point.
Robert.
Y
Yousuf Khan
Robert said:People who write to specific platforms will always have to count on
the continued availability and competitiveness of the platform. x86
is number one in both of those categories, and it's the only plausible
candidate to ride out the microprocessor revolution. I'd bet on it
again at this point.
What changed your mind? I thought Itanium was the only way to go,
according to you?
Yousuf Khan
R
Robert Myers
What changed your mind? I thought Itanium was the only way to go,
according to you?
I hope Itanium isn't the new Rambus--the thing you bring up when you
want to start a flame war.
Robert.
Y
Yousuf Khan
Robert said:I hope Itanium isn't the new Rambus--the thing you bring up when you
want to start a flame war.
Not at all, I've been convinced about x86 being here to stay for a long
time, while you have not. It's a legitimate question.
Regardless, OpenCL seems to promise a very flexible future, where you
split up workloads between all kinds of devices and not worry about what
they are.
Yousuf Khan
R
Robert Myers
Not at all, I've been convinced about x86 being here to stay for a long
time, while you have not. It's a legitimate question.
I"ve been around a long time. I've gotten some things right and some
things wrong.
Regardless, OpenCL seems to promise a very flexible future, where you
split up workloads between all kinds of devices and not worry about what
they are.
I'm skeptical of most meta-software. The world is a blizzard of
languages and API's, and software just gets to be more and more of a
mess. There's always some new whiz-bang thing that's going to save
the planet, or at least the industry.
Robert.
Y
Yousuf Khan
Robert said:I"ve been around a long time. I've gotten some things right and some
things wrong.
Well then, that's a very good answer.
I'm skeptical of most meta-software. The world is a blizzard of
languages and API's, and software just gets to be more and more of a
mess. There's always some new whiz-bang thing that's going to save
the planet, or at least the industry.
And according to the article, AMD wrote the OpenCL GPU drivers in a few
weeks. That's pretty much answering all of Nvidia's CUDA efforts, which
Nvidia has done over several years, in those few weeks. If things can be
done that quickly in OpenCL, then this meta-software might actually do
what it's promising.
Yousuf Khan
R
Robert Myers
Well then, that's a very good answer.
The cosmos will die a thermodynamic death before I hear a similar
admission from the latency-is-everything crowd.
Robert.