I wrote a particularly computationally intense GlowScript/WebGL
program, a visualization of a hard-sphere gas, with a continually
updated distribution of speeds that, averaged over time, approximates
the theoretical prediction:
It was rather sluggish compared to a similar VPython program, but the
VPython version benefited from the use of the fast parallel processing
capabilities of the Python numpy numerical library, which runs at the
speed of C, so the sluggishness was not totally unexpected.
But then to my great surprise I saw the new program running about
twice as fast on my Windows laptop as on my Windows desktop machine,
despite their comparable speeds (2.2 GHz desktop, 2.4 GHz laptop, and
the desktop has 4 cores, the laptop only 2). This seems to be related
to the fact that Chrome exists only in a 32-bit version, and my
desktop machine is 64-bit Windows 7 whereas my laptop is 32-bit
Windows 7. I thought naively that 32-bit programs ran on 64-bit
Windows at the same speed as they would run on 32-bit machines, but
maybe that isn't the case?
Strangely, on my desktop machine the program runs twice as fast in
Firefox as in Chrome, although my Firefox is also 32-bit. The program
also runs at this faster speed on my 2.4 GHz Macbook Pro in Safari.
The odd man out is Chrome on 64-bit Windows.
Odd, runs way too fast to pay attention to details with Chrome on Ubuntu
12.04 64bit, Lenovo X220 tablet quad core with Intel video, don't have any
alternatives to compare against.
Maybe we should take our laptops out to lunch at Second Street?
You have -- I can't resist the chemist's dig -- a physicist's calibration
of computationally intense. Try computing protein structures, or water
solvation shells around protein structures, or electron densities in
solvation shells and protein structures. Did I forget to add the ions?
Add the ions! Can't have decent structures without electrostatics.
I understand, intense relative to the other simulations, we'll get to to
density functional quantum simulations in their turn.
-- rec --
On Fri, Jun 8, 2012 at 8:32 PM, Bruce Sherwood <bruce.sherw...@gmail.com>wrote:
> I wrote a particularly computationally intense GlowScript/WebGL
> program, a visualization of a hard-sphere gas, with a continually
> updated distribution of speeds that, averaged over time, approximates
> the theoretical prediction:
> It was rather sluggish compared to a similar VPython program, but the
> VPython version benefited from the use of the fast parallel processing
> capabilities of the Python numpy numerical library, which runs at the
> speed of C, so the sluggishness was not totally unexpected.
> But then to my great surprise I saw the new program running about
> twice as fast on my Windows laptop as on my Windows desktop machine,
> despite their comparable speeds (2.2 GHz desktop, 2.4 GHz laptop, and
> the desktop has 4 cores, the laptop only 2). This seems to be related
> to the fact that Chrome exists only in a 32-bit version, and my
> desktop machine is 64-bit Windows 7 whereas my laptop is 32-bit
> Windows 7. I thought naively that 32-bit programs ran on 64-bit
> Windows at the same speed as they would run on 32-bit machines, but
> maybe that isn't the case?
> Strangely, on my desktop machine the program runs twice as fast in
> Firefox as in Chrome, although my Firefox is also 32-bit. The program
> also runs at this faster speed on my 2.4 GHz Macbook Pro in Safari.
> The odd man out is Chrome on 64-bit Windows.
> Bruce
> --
> You received this message because you are subscribed to the Santa Fe
> Complex "discuss" group.
> To post to this group, send email to disc...@sfcomplex.org
> To unsubscribe from this group, send email to
> discuss+unsubscr...@sfcomplex.org
> For more options, visit this group at
> http://groups.google.com/a/sfcomplex.org/group/discuss
Yes, by all means let's have dueling laptops at Second Street! When?
Of course you're right about "intense". Physicists also have a
calibration of what is really intense, e.g. quantum chromodynamics to
calculate from first principles the mass of the proton, or the
gravitational evolution of the universe starting with the Big Bang.
Nowadays much intense computation whether in chemistry or in physics
is done in GPUs, not CPUs.
GlowScript is a somewhat unusual project in that its goal is to make
it easy to program (albeit relatively non-intense) 3D visualizations
and animations, using WebGL. At least currently it doesn't give you
access to doing some of your computations in the GPU. In fact, you can
find unbelievably spectacular WebGL demos on the web, but each is a
special-purpose GPU-intensive implementation, whereas GlowScript is a
general-purpose tool.
On Sat, Jun 9, 2012 at 3:22 AM, Roger Critchlow <r...@elf.org> wrote:
> Odd, runs way too fast to pay attention to details with Chrome on Ubuntu
> 12.04 64bit, Lenovo X220 tablet quad core with Intel video, don't have any
> alternatives to compare against.
> Maybe we should take our laptops out to lunch at Second Street?
> You have -- I can't resist the chemist's dig -- a physicist's calibration of
> computationally intense. Try computing protein structures, or water
> solvation shells around protein structures, or electron densities in
> solvation shells and protein structures. Did I forget to add the ions? Add
> the ions! Can't have decent structures without electrostatics.
> I understand, intense relative to the other simulations, we'll get to to
> density functional quantum simulations in their turn.
> -- rec --
> On Fri, Jun 8, 2012 at 8:32 PM, Bruce Sherwood <bruce.sherw...@gmail.com>
> wrote:
>> I wrote a particularly computationally intense GlowScript/WebGL
>> program, a visualization of a hard-sphere gas, with a continually
>> updated distribution of speeds that, averaged over time, approximates
>> the theoretical prediction:
>> It was rather sluggish compared to a similar VPython program, but the
>> VPython version benefited from the use of the fast parallel processing
>> capabilities of the Python numpy numerical library, which runs at the
>> speed of C, so the sluggishness was not totally unexpected.
>> But then to my great surprise I saw the new program running about
>> twice as fast on my Windows laptop as on my Windows desktop machine,
>> despite their comparable speeds (2.2 GHz desktop, 2.4 GHz laptop, and
>> the desktop has 4 cores, the laptop only 2). This seems to be related
>> to the fact that Chrome exists only in a 32-bit version, and my
>> desktop machine is 64-bit Windows 7 whereas my laptop is 32-bit
>> Windows 7. I thought naively that 32-bit programs ran on 64-bit
>> Windows at the same speed as they would run on 32-bit machines, but
>> maybe that isn't the case?
>> Strangely, on my desktop machine the program runs twice as fast in
>> Firefox as in Chrome, although my Firefox is also 32-bit. The program
>> also runs at this faster speed on my 2.4 GHz Macbook Pro in Safari.
>> The odd man out is Chrome on 64-bit Windows.
>> Bruce
>> --
>> You received this message because you are subscribed to the Santa Fe
>> Complex "discuss" group.
>> To post to this group, send email to disc...@sfcomplex.org
>> To unsubscribe from this group, send email to
>> discuss+unsubscr...@sfcomplex.org
>> For more options, visit this group at
>> http://groups.google.com/a/sfcomplex.org/group/discuss
> --
> You received this message because you are subscribed to the Santa Fe Complex
> "discuss" group.
> To post to this group, send email to disc...@sfcomplex.org
> To unsubscribe from this group, send email to
> discuss+unsubscr...@sfcomplex.org
> For more options, visit this group at
> http://groups.google.com/a/sfcomplex.org/group/discuss
> Yes, by all means let's have dueling laptops at Second Street! When?
> Of course you're right about "intense". Physicists also have a
> calibration of what is really intense, e.g. quantum chromodynamics to
> calculate from first principles the mass of the proton, or the
> gravitational evolution of the universe starting with the Big Bang.
> Nowadays much intense computation whether in chemistry or in physics
> is done in GPUs, not CPUs.
> GlowScript is a somewhat unusual project in that its goal is to make
> it easy to program (albeit relatively non-intense) 3D visualizations
> and animations, using WebGL. At least currently it doesn't give you
> access to doing some of your computations in the GPU. In fact, you can
> find unbelievably spectacular WebGL demos on the web, but each is a
> special-purpose GPU-intensive implementation, whereas GlowScript is a
> general-purpose tool.
> Bruce
> On Sat, Jun 9, 2012 at 3:22 AM, Roger Critchlow <r...@elf.org> wrote:
> > Odd, runs way too fast to pay attention to details with Chrome on Ubuntu
> > 12.04 64bit, Lenovo X220 tablet quad core with Intel video, don't have
> any
> > alternatives to compare against.
> > Maybe we should take our laptops out to lunch at Second Street?
> > You have -- I can't resist the chemist's dig -- a physicist's
> calibration of
> > computationally intense. Try computing protein structures, or water
> > solvation shells around protein structures, or electron densities in
> > solvation shells and protein structures. Did I forget to add the ions?
> Add
> > the ions! Can't have decent structures without electrostatics.
> > I understand, intense relative to the other simulations, we'll get to to
> > density functional quantum simulations in their turn.
> > -- rec --
> > On Fri, Jun 8, 2012 at 8:32 PM, Bruce Sherwood <bruce.sherw...@gmail.com
> > wrote:
> >> I wrote a particularly computationally intense GlowScript/WebGL
> >> program, a visualization of a hard-sphere gas, with a continually
> >> updated distribution of speeds that, averaged over time, approximates
> >> the theoretical prediction:
> >> It was rather sluggish compared to a similar VPython program, but the
> >> VPython version benefited from the use of the fast parallel processing
> >> capabilities of the Python numpy numerical library, which runs at the
> >> speed of C, so the sluggishness was not totally unexpected.
> >> But then to my great surprise I saw the new program running about
> >> twice as fast on my Windows laptop as on my Windows desktop machine,
> >> despite their comparable speeds (2.2 GHz desktop, 2.4 GHz laptop, and
> >> the desktop has 4 cores, the laptop only 2). This seems to be related
> >> to the fact that Chrome exists only in a 32-bit version, and my
> >> desktop machine is 64-bit Windows 7 whereas my laptop is 32-bit
> >> Windows 7. I thought naively that 32-bit programs ran on 64-bit
> >> Windows at the same speed as they would run on 32-bit machines, but
> >> maybe that isn't the case?
> >> Strangely, on my desktop machine the program runs twice as fast in
> >> Firefox as in Chrome, although my Firefox is also 32-bit. The program
> >> also runs at this faster speed on my 2.4 GHz Macbook Pro in Safari.
> >> The odd man out is Chrome on 64-bit Windows.
> >> Bruce
> >> --
> >> You received this message because you are subscribed to the Santa Fe
> >> Complex "discuss" group.
> >> To post to this group, send email to disc...@sfcomplex.org
> >> To unsubscribe from this group, send email to
> >> discuss+unsubscr...@sfcomplex.org
> >> For more options, visit this group at
> >> http://groups.google.com/a/sfcomplex.org/group/discuss
> > --
> > You received this message because you are subscribed to the Santa Fe
> Complex
> > "discuss" group.
> > To post to this group, send email to disc...@sfcomplex.org
> > To unsubscribe from this group, send email to
> > discuss+unsubscr...@sfcomplex.org
> > For more options, visit this group at
> > http://groups.google.com/a/sfcomplex.org/group/discuss
> --
> You received this message because you are subscribed to the Santa Fe
> Complex "discuss" group.
> To post to this group, send email to disc...@sfcomplex.org
> To unsubscribe from this group, send email to
> discuss+unsubscr...@sfcomplex.org
> For more options, visit this group at
> http://groups.google.com/a/sfcomplex.org/group/discuss
On Sat, Jun 9, 2012 at 3:22 AM, Roger Critchlow <r...@elf.org> wrote:
> Odd, runs way too fast to pay attention to details with Chrome on Ubuntu
> 12.04 64bit, Lenovo X220 tablet quad core with Intel video, don't have any
> alternatives to compare against.
Awhile ago I did someWebGL tests on my iMac with different browsers (Chrome, Firefox, Safari) under both Mac OS X and Windows XP using
Bootcamp. The results were all over the place. For a given browser,
the JS engine differs depending on the OS as does the graphics
implementation. As I remember, the best performances were under Mac OS
X since OpenGL is the native graphics and under Windows the
browsers covert the OpenGL calls to Direct X. In addition. some of my online students
have had problems with 64-bit Windows drivers.
Ed
__________
Ed Angel
Founding Director, Art, Research, Technology and Science Laboratory (ARTS Lab)
Professor Emeritus of Computer Science, University of New Mexico
> I wrote a particularly computationally intense GlowScript/WebGL
> program, a visualization of a hard-sphere gas, with a continually
> updated distribution of speeds that, averaged over time, approximates
> the theoretical prediction:
> It was rather sluggish compared to a similar VPython program, but the
> VPython version benefited from the use of the fast parallel processing
> capabilities of the Python numpy numerical library, which runs at the
> speed of C, so the sluggishness was not totally unexpected.
> But then to my great surprise I saw the new program running about
> twice as fast on my Windows laptop as on my Windows desktop machine,
> despite their comparable speeds (2.2 GHz desktop, 2.4 GHz laptop, and
> the desktop has 4 cores, the laptop only 2). This seems to be related
> to the fact that Chrome exists only in a 32-bit version, and my
> desktop machine is 64-bit Windows 7 whereas my laptop is 32-bit
> Windows 7. I thought naively that 32-bit programs ran on 64-bit
> Windows at the same speed as they would run on 32-bit machines, but
> maybe that isn't the case?
> Strangely, on my desktop machine the program runs twice as fast in
> Firefox as in Chrome, although my Firefox is also 32-bit. The program
> also runs at this faster speed on my 2.4 GHz Macbook Pro in Safari.
> The odd man out is Chrome on 64-bit Windows.
> Bruce
> -- > You received this message because you are subscribed to the Santa Fe Complex "discuss" group.
> To post to this group, send email to disc...@sfcomplex.org
> To unsubscribe from this group, send email to
> discuss+unsubscr...@sfcomplex.org
> For more options, visit this group at
> http://groups.google.com/a/sfcomplex.org/group/discuss
32 bit .deb (For Debian/Ubuntu)
64 bit .deb (For Debian/Ubuntu) 32 bit .rpm (For Fedora/openSUSE) 64 bit
.rpm (For Fedora/openSUSE)
I'm running the 64 bit .deb (For Debian/Ubuntu), haven't looked at it too
closely to see what exactly is in it.
-- rec --
On Sat, Jun 9, 2012 at 10:13 AM, Bruce Sherwood <bruce.sherw...@gmail.com>wrote:
> I read somewhere that there is a 64-bit version of Chrome available
> only on Linux. Maybe that's what you're running?
> Bruce
> On Sat, Jun 9, 2012 at 3:22 AM, Roger Critchlow <r...@elf.org> wrote:
> > Odd, runs way too fast to pay attention to details with Chrome on Ubuntu
> > 12.04 64bit, Lenovo X220 tablet quad core with Intel video, don't have
> any
> > alternatives to compare against.
> --
> You received this message because you are subscribed to the Santa Fe
> Complex "discuss" group.
> To post to this group, send email to disc...@sfcomplex.org
> To unsubscribe from this group, send email to
> discuss+unsubscr...@sfcomplex.org
> For more options, visit this group at
> http://groups.google.com/a/sfcomplex.org/group/discuss
