Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Supercomputer Breaks the $100/GFLOPS Barrier

Author: Robert Hyatt

Date: 21:31:23 08/27/03

Go up one level in this thread


On August 28, 2003 at 00:17:52, Tom Kerrigan wrote:

>On August 27, 2003 at 17:26:21, Robert Hyatt wrote:
>
>>On August 27, 2003 at 00:02:32, Tom Kerrigan wrote:
>>
>>>On August 25, 2003 at 19:56:00, Robert Hyatt wrote:
>>>
>>>>On August 25, 2003 at 06:13:58, Bo Persson wrote:
>>>>
>>>>>On August 25, 2003 at 03:36:03, Russell Reagan wrote:
>>>>>
>>>>>>On August 25, 2003 at 01:43:23, Sally Weltrop wrote:
>>>>>>
>>>>>>>On August 24, 2003 at 16:17:07, Sietel Monic wrote:
>>>>>>>
>>>>>>>>>yeap, been with them for years now. If it was just intel alone we would be
>>>>>>>>>paying $1,500 for a 3GHz chip. My next machine is going to be a Athlon Dual
>>>>>>>>>3.2GHz
>>>>>>>>
>>>>>>>>Amd is coming out with dual athlon Mp doing 3.2 ghz(each?)? Or is this both
>>>>>>>>combined processor speeds?
>>>>>>>
>>>>>>>
>>>>>>>http://speedycpu.dyndns.org/mp3200.htm
>>>>>>
>>>>>>That is 2.4GHz, not 3.2GHz. Yes, the 3200 is misleading, probably on purpose :)
>>>>>
>>>>>The 3.2 vs 2.4 is also misleading. Definitely on purpose.
>>>>>
>>>>>Look - my engine is 200 HP at 3200 rpm, his is just 2400 rpm...
>>>>>
>>>>>200 HP V8
>>>>>200 HP V6
>>>>>200 HP V4 Turbo
>>>>>
>>>>>Which car runs the fastest?   :-)
>>>>
>>>>I'll bet the V8 wins the drag race..   (hint:  torque)
>>>>
>>>>:)
>>>
>>>[Effective] torque depends on the engine and gearing. You can't answer the
>>>question without knowing how the cars are geared.
>>>
>>>Also, turbos usually give you a very broad range of peak torque. I would expect
>>>a "typical" turbo I4 (not V4) car to win a drag race vs. a typical V8. (Assuming
>>>same HP.)
>>>
>>>-Tom
>>
>>As a general rule, gearing is irrelevant in the comparison.  Whatever gearset
>>you run in the v6, I'll run in the v8 for the 1/4 mile.  And the v8 is going to
>>eat the v6.  From experience.  Because whatever RPM you turn in the v6, I can
>>get from the v8.  So whatever torque multiplication you get with your 4.10 (or
>>whatever) rear end, I will also get in the v8.
>
>Well, as the saying goes, you can either have high displacement or high RPMs but
>not both... unless you're Ferrari. How many I4s redline at 8000-9000 these days,
>and how many V8s under $100,000 redline in that range?

Not many stock motors that appear in _any_ vehicle redline up there.  IE,
for comparison, my son's SHO taurus redlines at about 7K.  The current 4.6
ford V8 redlines in the same area.  As did the SVO I4 with turbo.

A bit more exotic is the Honda S2000 at near 9K, but it has no torque,
after driving one.


>
>>We are talking pure motor issues here.  Horsepower is important.  But so is
>>torque, which is a different way of measuring horsepower.  8 cylinders will
>>produce more torque, even at the same displacement.  A V10 is better still.
>
>That may generally be the case, but torque technically depends on the shape of
>the cylinders. I'm sure it's easy to make a V6 that has more torque than a V8
>with the same displacement.

No idea what you are talking about when you say "shape of the cylinders."  The
_only_ shape being used is a perfect circle of varying radius, with a taller
or shorter swept area controlled by the stroke of the crank.

However, it is definitely impossible to make more torque with fewer cylinders,
if all else is the same.  Just compare the difference between a power pulse
every 1/4 revolution vs every 1/3 revolution.  That's why v12's exist.  More
torque.  Smoother since power pulses are closer together in terms of 1/6 of
a revolution between power pulses vs 1/4 or 1/3...


>
>>The problem with the turbo is lag.  The V8 supplies torque _right now_.  The
>>I4 is going to die at low RPM, and waits until the turbo spools up to get
>>moving.  A good example is the old SVO mustang with an I4.  Yes, it runs like
>
>Which is fine, because if you're drag racing you can spool up the turbo before
>launching, so turbo lag is not an issue unless your car is so poorly tuned that
>you experience lag after you upshift at redline. Turbo lag is only significant
>for rolling starts and actual race courses.

The turbo won't spool up.  It depends on exhaust volume.  A free-wheeling
engine won't produce the exhaust gas volume that an engine under load
produces, you simply don't pump as much air through the engine, nor as much
gas, which translates into less exhaust volume.


>
>>the devil.  Yes, a good 302 will suck the doors off of it.  HP is about the same
>>(220 each).  My son and a friend used to prove this over and over.  My son had
>>the 302, his friend had a mint SVO.  The SVO would run for an I4 (13.7 quarter
>>miles.)  The 302 ran 13.3's.
>
>That's great, but the gearing could have been different, the turbo could have
>been tuned poorly for the 1/4, and the friend might not have been launching
>right. Actually, unless he had a race clutch, I doubt he was launching right.

This was a regular drag racer, using a comp clutch and pressure plate, using
drag radial tires.  He knew what he was doing.  I drove the car myself and we
were within .1 of each other...

>
>>But the main test is to start at rest and see what that turbo does.  My son
>>now has an SHO taurus with a v6, and yes, it runs great.  Same horsepower as
>
>Are you sure? I thought all SHO Tauri were V8s.
>

Absolutely positive.   Motor built by Yamaha.  6 cylinder.  3.0 liters.  220
HP.  (1995 model year).,

Got the dual-runner intake that opens short runners at 4500 RPM.  You could
drive all three cars (same factory HP rating at 220) and there was no
comparison in the v8 vs the others.  The SVO 4 cylinder has more torque than
the unblown v6, but that is no surprise, as the I4 was 2.4 liters (I think)
while the V6 is a 3.0.  But the 5.0 V8 just blows them both off on the bottom
end.  The 5.0 has run 13.0's with real 26 x 10 slicks.  It has enough torque
that it needs more rubber.  The SVO turbo and the SHO don't have that kind of
traction problem...

>-Tom



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.