OT: Wheeeeee! New PSpice Benchmarks

[Jim Thompson]

I just got my new 2.2GHz AMD Athlon 64 machine up and running
(replacement for the 1.467GHz AMD Athlon that died last week).

Ran the benchmark I posted 3 years ago.

2.175 times faster !!!

...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice480)460-2350 | |
| E-mail Address at Website Fax480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

 

[Active8]

On Sun, 22 Aug 2004 16:54:06 -0700, Jim Thompson wrote:

I just got my new 2.2GHz AMD Athlon 64 machine up and running
(replacement for the 1.467GHz AMD Athlon that died last week).

Ran the benchmark I posted 3 years ago.

2.175 times faster !!!

...Jim Thompson

Refresh my DRAM. What was the subject line? I don't think I've got a
benchmark.
--
Best Regards,
Mike

 

[Jim Thompson]

On Sun, 22 Aug 2004 20:05:46 -0400, Active8 <reply2group@ndbbm.net>
wrote:

On Sun, 22 Aug 2004 16:54:06 -0700, Jim Thompson wrote:

I just got my new 2.2GHz AMD Athlon 64 machine up and running
(replacement for the 1.467GHz AMD Athlon that died last week).

Ran the benchmark I posted 3 years ago.

2.175 times faster !!!

...Jim Thompson

Refresh my DRAM. What was the subject line? I don't think I've got a
benchmark.

On 8/23/2001 I posted that a benchmark circuit was available for those
who wanted to try it.

You had to request it by E-mail, since I didn't want my design in just
any old hands ;-)

Three years ago, when I first posted the results, my fastest machine
was an 800MHz P3... the new Athlon simulates at 5.28X faster.

Here are the tabulated results.

Computer SIM Time, seconds
=========== =================
P2 266MHz 1800
P2 440MHz 1082
P3 800MHz 510
P3 1000MHz 420
P4 1500MHz 413
ATH 1.2GHz 270/244 depending on motherboard
ATH 1.4GHz 210
ATH 2.2GHz 96.58

One thing is readily apparent... P4s suck at mathematically intensive
tasks... that's why I switched to AMD Athlons three years ago.

...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice480)460-2350 | |
| E-mail Address at Website Fax480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

 

[Active8]

On Sun, 22 Aug 2004 17:32:49 -0700, Jim Thompson wrote:

On Sun, 22 Aug 2004 20:05:46 -0400, Active8 <reply2group@ndbbm.net
wrote:

On Sun, 22 Aug 2004 16:54:06 -0700, Jim Thompson wrote:

I just got my new 2.2GHz AMD Athlon 64 machine up and running
(replacement for the 1.467GHz AMD Athlon that died last week).

Ran the benchmark I posted 3 years ago.

2.175 times faster !!!

...Jim Thompson

Refresh my DRAM. What was the subject line? I don't think I've got a
benchmark.

On 8/23/2001 I posted that a benchmark circuit was available for those
who wanted to try it.

You had to request it by E-mail, since I didn't want my design in just
any old hands ;-)

Three years ago, when I first posted the results, my fastest machine
was an 800MHz P3... the new Athlon simulates at 5.28X faster.

Here are the tabulated results.

Computer SIM Time, seconds
=========== =================
P2 266MHz 1800
P2 440MHz 1082
P3 800MHz 510
P3 1000MHz 420
P4 1500MHz 413
ATH 1.2GHz 270/244 depending on motherboard
ATH 1.4GHz 210
ATH 2.2GHz 96.58

One thing is readily apparent... P4s suck at mathematically intensive
tasks... that's why I switched to AMD Athlons three years ago.

...Jim Thompson

So did my email addy make it into your whitelist?
--
Best Regards,
Mike

 

[Active8]

On Mon, 23 Aug 2004 00:59:07 +0000 (UTC), Ken Smith wrote:

In article <t2eii0dkt5m3guqegeb1k3p6mehblt9oms@4ax.com>,
Jim Thompson <thegreatone@example.com> wrote:
[...]
Computer SIM Time, seconds
=========== =================
P2 266MHz 1800
P2 440MHz 1082
P3 800MHz 510
P3 1000MHz 420
P4 1500MHz 413
ATH 1.2GHz 270/244 depending on motherboard
ATH 1.4GHz 210
ATH 2.2GHz 96.58

1800*266/1500=319.2 so clock cycle for clock cycle Intel has lost ground.

244*1.2/2.2=133.1 AMD seems to be making better use of clock cycles.

I assume these were all with the same OS.

Good point, but the people I know that build gaming systems all go
with AMD. The graphics are math intensive. They all tell me that the
AMDs do the job faster for a civen CPU speed.
--
Best Regards,
Mike

 

[Chaos Master]

It was written by Jim Thompson[thegreatone@example.com] in message
<b5cii09ivp9fvpton296dvu1n7quggig40@4ax.com>:

I just got my new 2.2GHz AMD Athlon 64 machine up and running
(replacement for the 1.467GHz AMD Athlon that died last week).

Ran the benchmark I posted 3 years ago.

2.175 times faster !!!

What is the benchmark circuit?

I am interested in running it on my Bochs [1] virtual machine that is running
Windows 98 -- to try to discover emulation bugs.

[1] http://bochs.sf.net , free 'virtual machine' program (allows to run OS
inside OS. runs almost all OS'es, from DOS 3.3 to Linux)


--
Chaos MasterŽ, posting from Brazil. REPLY TO GROUP!
MSN: wizard_of_yendor[@]hotmail[.]com http://marreka.no-ip.com
"People told me I can't dress like a fairy. I say,
I'm in a rock band and I can do what the hell I want!" -- Amy Lee

 

[Rich Grise]

Jim Thompson wrote:

I just got my new 2.2GHz AMD Athlon 64 machine up and running
(replacement for the 1.467GHz AMD Athlon that died last week).

Ran the benchmark I posted 3 years ago.

2.175 times faster !!!

What are your bogomips?
;-)

 

[Jim Thompson]

On Sun, 22 Aug 2004 20:58:16 -0400, Active8 <reply2group@ndbbm.net>
wrote:

On Sun, 22 Aug 2004 17:32:49 -0700, Jim Thompson wrote:

On Sun, 22 Aug 2004 20:05:46 -0400, Active8 <reply2group@ndbbm.net
wrote:

On Sun, 22 Aug 2004 16:54:06 -0700, Jim Thompson wrote:

I just got my new 2.2GHz AMD Athlon 64 machine up and running
(replacement for the 1.467GHz AMD Athlon that died last week).

Ran the benchmark I posted 3 years ago.

2.175 times faster !!!

...Jim Thompson

Refresh my DRAM. What was the subject line? I don't think I've got a
benchmark.

On 8/23/2001 I posted that a benchmark circuit was available for those
who wanted to try it.

You had to request it by E-mail, since I didn't want my design in just
any old hands ;-)

Three years ago, when I first posted the results, my fastest machine
was an 800MHz P3... the new Athlon simulates at 5.28X faster.

Here are the tabulated results.

Computer SIM Time, seconds
=========== =================
P2 266MHz 1800
P2 440MHz 1082
P3 800MHz 510
P3 1000MHz 420
P4 1500MHz 413
ATH 1.2GHz 270/244 depending on motherboard
ATH 1.4GHz 210
ATH 2.2GHz 96.58

One thing is readily apparent... P4s suck at mathematically intensive
tasks... that's why I switched to AMD Athlons three years ago.

...Jim Thompson

So did my email addy make it into your whitelist?

Go to the website and use that address, and then I'll respond.

(That address will be valid for awhile... I'm changing over to a
version of formmail, since some schmuck has submitted the *image-only*
address on the website to spammers.)

...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice480)460-2350 | |
| E-mail Address at Website Fax480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

 

[Jim Thompson]

On Mon, 23 Aug 2004 00:59:07 +0000 (UTC), kensmith@green.rahul.net
(Ken Smith) wrote:

In article <t2eii0dkt5m3guqegeb1k3p6mehblt9oms@4ax.com>,
Jim Thompson <thegreatone@example.com> wrote:
[...]
Computer SIM Time, seconds
=========== =================
P2 266MHz 1800
P2 440MHz 1082
P3 800MHz 510
P3 1000MHz 420
P4 1500MHz 413
ATH 1.2GHz 270/244 depending on motherboard
ATH 1.4GHz 210
ATH 2.2GHz 96.58

1800*266/1500=319.2 so clock cycle for clock cycle Intel has lost ground.

244*1.2/2.2=133.1 AMD seems to be making better use of clock cycles.

I assume these were all with the same OS.


--

Yep, Win2K Pro.

...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice480)460-2350 | |
| E-mail Address at Website Fax480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

 

[Jim Thompson]

On Sun, 22 Aug 2004 22:38:42 -0300, Chaos Master
<chaos.master@yahoo.com> wrote:

in message
b5cii09ivp9fvpton296dvu1n7quggig40@4ax.com>:
I just got my new 2.2GHz AMD Athlon 64 machine up and running
(replacement for the 1.467GHz AMD Athlon that died last week).

Ran the benchmark I posted 3 years ago.

2.175 times faster !!!

What is the benchmark circuit?

I am interested in running it on my Bochs [1] virtual machine that is running
Windows 98 -- to try to discover emulation bugs.

[1] http://bochs.sf.net , free 'virtual machine' program (allows to run OS
inside OS. runs almost all OS'es, from DOS 3.3 to Linux)

Go to the website and use that address to contact me, and then I'll
respond.

(That address will be valid for awhile... I'm changing over to a
version of formmail, since some schmuck has submitted the *image-only*
address on the website to spammers.)

...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice480)460-2350 | |
| E-mail Address at Website Fax480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

 

[Jim Thompson]

On Mon, 23 Aug 2004 02:11:47 +0000 (UTC), kensmith@green.rahul.net
(Ken Smith) wrote:

In article <b2wqyl86ltmc$.dlg@news.individual.net>,
Active8 <reply2group@ndbbm.net> wrote:
[...]
Good point, but the people I know that build gaming systems all go
with AMD. The graphics are math intensive. They all tell me that the
AMDs do the job faster for a civen CPU speed.

We don't know that "A" is the cause of "B" in this case. 3D graphics
requires fast low accuracy math and a bunch of memory transfers that don't
simply count up. The AMD may do both better.

It may not even be the CPU core its self that is the cause. These chips
have cache memories built in. Most cache designs assume things about the
pattern of reads and writes. If it assumes wrong, the performance
suffers.

--

When Intel went to the P4 they tossed the dedicated math core, so they
could devote more chip area to the pablum that the general public
wants.

...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice480)460-2350 | |
| E-mail Address at Website Fax480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

 

[Ken Smith]

In article <t2eii0dkt5m3guqegeb1k3p6mehblt9oms@4ax.com>,
Jim Thompson <thegreatone@example.com> wrote:
[...]

Computer SIM Time, seconds
=========== =================
P2 266MHz 1800
P2 440MHz 1082
P3 800MHz 510
P3 1000MHz 420
P4 1500MHz 413
ATH 1.2GHz 270/244 depending on motherboard
ATH 1.4GHz 210
ATH 2.2GHz 96.58

1800*266/1500=319.2 so clock cycle for clock cycle Intel has lost ground.

244*1.2/2.2=133.1 AMD seems to be making better use of clock cycles.

I assume these were all with the same OS.


--
--
kensmith@rahul.net forging knowledge

 

[Active8]

On Sun, 22 Aug 2004 22:38:42 -0300, Chaos Master wrote:

in message
b5cii09ivp9fvpton296dvu1n7quggig40@4ax.com>:
I just got my new 2.2GHz AMD Athlon 64 machine up and running
(replacement for the 1.467GHz AMD Athlon that died last week).

Ran the benchmark I posted 3 years ago.

2.175 times faster !!!

What is the benchmark circuit?

I am interested in running it on my Bochs [1] virtual machine that is running
Windows 98 -- to try to discover emulation bugs.

[1] http://bochs.sf.net , free 'virtual machine' program (allows to run OS
inside OS. runs almost all OS'es, from DOS 3.3 to Linux)

Do you know if and why that VM is better than VMWare? Aside from the
fact that VMWare went from free to not free, save that obscure link
from FreeBSD to the older versions.
--
Best Regards,
Mike

 

[Mike Engelhardt]

Jim,

Ran the benchmark I posted 3 years ago.

2.175 times faster !!!

But isn't that the benchmark that LTspice ran about 2.5x
faster than PSpice? Could have gotten 2 years and 3
months on Moore's Law just by switching to LTspice.

--Mike

 

[KR Williams]

In article <cgbfgr$83b$2@blue.rahul.net>,
kensmith@green.rahul.net says...

In article <t2eii0dkt5m3guqegeb1k3p6mehblt9oms@4ax.com>,
Jim Thompson <thegreatone@example.com> wrote:
[...]
Computer SIM Time, seconds
=========== =================
P2 266MHz 1800
P2 440MHz 1082
P3 800MHz 510
P3 1000MHz 420
P4 1500MHz 413
ATH 1.2GHz 270/244 depending on motherboard
ATH 1.4GHz 210
ATH 2.2GHz 96.58

1800*266/1500=319.2 so clock cycle for clock cycle Intel has lost ground.

244*1.2/2.2=133.1 AMD seems to be making better use of clock cycles.

P4s are well known to be crappy processors with respect to IPC.
The P3 was respectable, but the P4 went for the MHz, performance
be damned. To be fair, if you're doing DV (and only DVish
things), the P4 is a respectable processor.

The original P4's "forgot" the integer multiplier and the barrel
shifter. Each is a severe performance hit for fixed point
operations. To do integer multiplies the data must be sent
across the chip *twice* (over and back) to the FPU, which ties up
the FPU as well. Rumor has it that they've sorta fixed this issue
in the later chips, but no one is really saying.

BTW, I went with the Opteron 144 (1.8GHz) because it has a 1MB
cache (as opposed to the A64's 512K). ...couldn't afford a G5.

I assume these were all with the same OS.

I'd hope we're talking a 64b OS and software for the A64.
Otherwise the gains are truly remarkable!

--
Keith

 

[KR Williams]

In article <cgbjp3$9fi$1@blue.rahul.net>,
kensmith@green.rahul.net says...

In article <b2wqyl86ltmc$.dlg@news.individual.net>,
Active8 <reply2group@ndbbm.net> wrote:
[...]
Good point, but the people I know that build gaming systems all go
with AMD. The graphics are math intensive. They all tell me that the
AMDs do the job faster for a civen CPU speed.

We don't know that "A" is the cause of "B" in this case. 3D graphics
requires fast low accuracy math and a bunch of memory transfers that don't
simply count up. The AMD may do both better.

The GPU does most of the "low accuracy" math in gamer systems.
Both AMD and Intel processors have compatible arithmetic
instruction sets.

It may not even be the CPU core its self that is the cause. These chips
have cache memories built in. Most cache designs assume things about the
pattern of reads and writes. If it assumes wrong, the performance
suffers.

It's not so much the caches (though that matters, and AMD's
exclusivity may be somewhat important here) as the memory
interface. AMD has it all over Intel with their integrated DRAM
controllers. Bandwidth may be expensive (buy an Opteron), but
latency is forever.

--
Keith

 

[Chaos Master]

It was written by Jim Thompson[thegreatone@example.com] in message
<u2mii0d21jc2i6dq08trcsk2b0rbjpovg9@4ax.com>:

What is the benchmark circuit?

I am interested in running it on my Bochs [1] virtual machine that is running
Windows 98 -- to try to discover emulation bugs.

[1] http://bochs.sf.net , free 'virtual machine' program (allows to run OS
inside OS. runs almost all OS'es, from DOS 3.3 to Linux)

Go to the website and use that address to contact me, and then I'll
respond.

(That address will be valid for awhile... I'm changing over to a
version of formmail, since some schmuck has submitted the *image-only*
address on the website to spammers.)

I have sent one e-mail. My e-mail is

renan_tdb @@ yahoo .. com .. br (remove one '@' and one '.' in each '..' combo)

[]s
--
Chaos MasterŽ, posting from Brazil. REPLY TO GROUP!
MSN: wizard_of_yendor[@]hotmail[.]com http://marreka.no-ip.com
"People told me I can't dress like a fairy. I say,
I'm in a rock band and I can do what the hell I want!" -- Amy Lee

 

[Chaos Master]

It was written by Active8[reply2group@ndbbm.net] in message
<pc5gt7ftk673.dlg@news.individual.net>:

What is the benchmark circuit?

I am interested in running it on my Bochs [1] virtual machine that is running
Windows 98 -- to try to discover emulation bugs.

[1] http://bochs.sf.net , free 'virtual machine' program (allows to run OS
inside OS. runs almost all OS'es, from DOS 3.3 to Linux)

Do you know if and why that VM is better than VMWare? Aside from the
fact that VMWare went from free to not free, save that obscure link
from FreeBSD to the older versions.

I like Bochs, mainly because:

1. it's free (open source),
2. runs in lots of platforms/processors (since it emulates a total x86 it can
run on any Windows (98 and above), Mac OS X, Linux, SGI IRIX, AIX... while
VMWare only runs on x86 Windows (NT/2000/XP) /Linux. I have a friend -
Macintosh user - that uses Bochs with Windows to play some games)
3. It has a debugger (not that I need it, but...)


And it's slow, but it's OK to run some old software.
I use Windows (1.0, 3.1, 3.11) and DOS (3.3, 6.22) inside it.
Better than in my old 386, I can say.


[]s
--
Chaos MasterŽ, posting from Brazil. REPLY TO GROUP!
MSN: wizard_of_yendor[@]hotmail[.]com http://marreka.no-ip.com
"People told me I can't dress like a fairy. I say,
I'm in a rock band and I can do what the hell I want!" -- Amy Lee

 

[Ken Smith]

In article <b2wqyl86ltmc$.dlg@news.individual.net>,
Active8 <reply2group@ndbbm.net> wrote:
[...]

Good point, but the people I know that build gaming systems all go
with AMD. The graphics are math intensive. They all tell me that the
AMDs do the job faster for a civen CPU speed.

We don't know that "A" is the cause of "B" in this case. 3D graphics
requires fast low accuracy math and a bunch of memory transfers that don't
simply count up. The AMD may do both better.

It may not even be the CPU core its self that is the cause. These chips
have cache memories built in. Most cache designs assume things about the
pattern of reads and writes. If it assumes wrong, the performance
suffers.

--
--
kensmith@rahul.net forging knowledge

 

[John S. Dyson]

In article <o4mii016pkf1gjno7i9d6u4f04au4ipifg@4ax.com>,
Jim Thompson <thegreatone@example.com> writes:

On Mon, 23 Aug 2004 02:11:47 +0000 (UTC), kensmith@green.rahul.net
(Ken Smith) wrote:

In article <b2wqyl86ltmc$.dlg@news.individual.net>,
Active8 <reply2group@ndbbm.net> wrote:
[...]
Good point, but the people I know that build gaming systems all go
with AMD. The graphics are math intensive. They all tell me that the
AMDs do the job faster for a civen CPU speed.

We don't know that "A" is the cause of "B" in this case. 3D graphics
requires fast low accuracy math and a bunch of memory transfers that don't
simply count up. The AMD may do both better.

It may not even be the CPU core its self that is the cause. These chips
have cache memories built in. Most cache designs assume things about the
pattern of reads and writes. If it assumes wrong, the performance
suffers.

--

When Intel went to the P4 they tossed the dedicated math core, so they
could devote more chip area to the pablum that the general public
wants.

The whole AMD vs. Intel argument isn't really 'interesting' to me,

but one thing about the IntelP4 -- it does tend to underperform
when using the old stack-style instructions for FP math. When using
the SSE/SSE2 instruction set, it performs much better, but with a slight
hit WRT math accuracy. My application includes FIR filtering, some
transforms and some matrix operations.

I have been writing significant amounts of FP code with the SSE
instruction set, and the P4 is very very fast. It is important
to take advantage of the SIMD capabilities (which isn't always possible)
instead of limping along with the 'stack' instructions. (Of course, the
stack instructions can give higher precision or give access to the
microcoded transcendental math at high accuracy.)

When doing DP math when using the stack instructions (which is what
most compilers tend to use) and comparing the SSE2 SIMD instructions,
based upon real code, the P4 can blow itself away (speed wise) in SSE2 mode.
2X performance improvement for vector (or matrix) operations for SSE2
operations isn't unrealistic (but that much isn't guaranteed.)

The SSE2 instructions do their best for multiple (SIMD) operations -- things
like array operations, which are great for DSP, vector and other
operations.

I'd still like to have an SMP FX53 Athlon64, mostly for the potential
for improved performance for sundry things (in 64 bit mode, some operations
appear to be very quick.) Normally,
I prefer dual processor machines, and am disappointed that upgrading
(or side-grading) to AMD64 from Intel would cause me to loose the ability
to simulate SMP multithreading when choosing the highest performing chips.
(My purposes for using SMP isn't always for max performance -- even though
the extra performance can be nice, but it is the ability to check out
code for real SMP machines... So, HTT helps to give me some of what I
want for SMP, even though it doesn't give as much extra performance as
SMP.) I'd happily 'upgrade' from IntelHTT to Athlon64 SMP, but would
loose the choice of the fastest machines for SMP (or SMP simulation.)

John