Just to know...

16 posts / 0 new
Last post
coius's picture
Offline
Last seen: 10 years 2 weeks ago
Joined: Aug 25 2004 - 13:56
Posts: 1975
Just to know...

Ok, I have a few questions.

first let me describe what I have. This will be the basis for all the questions.
This is the setup:

I have a computer with an MSI Motherboard (MS-5169) With Aladdin 5 Chipset Chipset (I think this is right)
The board accepts PII Or Pentium MMX and AMD K6/K6-2/III Processors. Depending on which you want to use. (it sports a Socket 7 Surface mount ZIF) Supports 2x AGP (see video card below for capabilities)
The board has onboard:

Ports:

1x Parallel Port (supports SPP/EPP/ECP Mode
2x IDE 33MB's Ports
1x Floppy port capable of supporting 2 FDD with 360k, 720k, 1.2MB, 1.44MB, ans 2.88MB Disks (anthing under the 1.2MB is most likely 5.25" and could be 3.5" as well)
2x Serial Ports (COM-A + COM-B)
4x 32-bit Master PCI Bus Slots and 3 ISA Bus Slots (support for 3.3v/5v PCI cards)
2x AGP (Correct me on the speed if there is a conflict) 66/133MHz 3.3v)

Ram:

3 slots Memory capable of 256MB in each slot (Up to 768MB Total)
Ram supports 3.3v EDO and SDRAM Dimms

Bios:
Upgraded to latest Rev. Plug & Play
Supports Soft Power-off
Alarm Wake-up

Processor Performance:
Support 512k Pipelined Cache Memory.
Cacheable Area: 128MB
The Processor is a 400MHz AMD K6-2 Clocked to 450MHz w/ 1MB Cache

Slots
PCI:
1x Yamaha Sound Card (Sports Hardware MIDI Synthesizer)
1x USB 1.1 2 port card
1x RealTek 10/100Mb's Ethernet LAN Card
1x LinkSys 11Mb's WIFI Card

AGP:
Trident Blade 3D - 4MB VRAM (not enough, I know) Hits 1280x 1024 Rez. at High Color (16-bit)
all other Rez can use True Color.

ISA
slot 1:
Nothing
slot 2:
Nothing

I have a 40x CD-rom in 1 of the Four 5.25" bay
the Hard Drive is an 80GB Seagate
I have 2 1.44MB Floppy drives hooked up to the FDD Controller

Question 1:
What is the Approximate worth of this system.

Question 2:
What is the equivelent Macintosh to this (NOT PRICE, Speed)

Question 4:
What is this related to an Intel system (P-1 [w w/o MMX?] / P-II) and the speed of the equivelent.

I only work on the systems and don't bother with the specs. This time I am a little curious.

What do you think?

Eudimorphodon's picture
Offline
Last seen: 3 months 3 weeks ago
Joined: Dec 21 2003 - 14:14
Posts: 1207
Re: Just to know...


Question 1:
What is the Approximate worth of this system.

"Nothing! Absolutely Nothing!" ;^>

(Or at least not much more then that, depending on where you're trying to unload it. The 80GB hard drive is probably "worth" as much as the rest of the system put together. Figure getting *maybe* $100 for the whole thing at a swap meet.

I've gotten more powerful computers for free. And given them away, so far as that goes.)

Question 2:
What is the equivelent Macintosh to this (NOT PRICE, Speed)

It'd be a pretty even match for a B&W G3. (Don't let a PowerPC fanatic tell you otherwise.)


Question 4:
What is this related to an Intel system (P-1 [w w/o MMX?] / P-II) and the speed of the equivelent.

Intel only sold desktop Pentium/PentiumMMX CPUs up to 233Mhz, so obviously it's faster then anything there. (At floating point operations the original Pentium is slightly faster clock-for-clock then K6, but you're running twice as fast, so it's not making up the difference.) Compared to a PII, well... in integer operations the K6 is faster. They used to clean up at Linux kernel compilation benchmarks, for instance. For sheer number-crunching oomph a PII at the same speed would probably clean your clock.

Compared to a brand new Pentium 4 or Athlon system, it's probably, oh, 1/10th as fast, depending on what you're doing.

Just for trivia's sake, I'll note that Applefritter.com was running on a system close to identical to this one for about two years after "Deep Dish" died. Said system was plenty fast for plain-old web serving, but keeled over when "Drupal" was installed on it.

--Peace

aladds's picture
Offline
Last seen: 1 month 2 days ago
Joined: Dec 26 2003 - 16:21
Posts: 298
Re: Just to know...

Question 2:
What is the equivelent Macintosh to this (NOT PRICE, Speed)

It'd be a pretty even match for a B&W G3. (Don't let a PowerPC fanatic tell you otherwise.)

I'd say about Powermac 9600/350 speed

Eudimorphodon's picture
Offline
Last seen: 3 months 3 weeks ago
Joined: Dec 21 2003 - 14:14
Posts: 1207
Re: Just to know...

I'd say about Powermac 9600/350 speed

To add a bit of weight to my estimation... ;^>

A couple years ago I ran an old C floating-point benchmark on some random machines I had access to.

Here's the results.

The 500Mhz K6-II slotted right between a 233Mhz iMac and a 600Mhz iBook. Which makes it look like a 350Mhz-ish G3 to me.

Since we're talking about a 400Mhz K6-II here, based on these results I'd say we're roughly in the 266Mhz-300Mhz G3 ballpark. Since we're on the cusp I'd go ahead and specify *Blue* G3 instead of Beige, since the 100Mhz bus, UDMA/33 IDE controller, and AGP video all substantially outspec the Beige variety.

--Peace

Offline
Last seen: 18 years 2 months ago
Joined: Jan 23 2005 - 20:28
Posts: 159
there is no way that is equal

there is no way that is equal to a b&w.probably at most a 9600/300-350.

Eudimorphodon's picture
Offline
Last seen: 3 months 3 weeks ago
Joined: Dec 21 2003 - 14:14
Posts: 1207
Re: there is no way that is equal

there is no way that is equal to a b&w.probably at most a 9600/300-350.

Is this because you think so, or because you have some sort of evidence to back that up?

--Peace

coius's picture
Offline
Last seen: 10 years 2 weeks ago
Joined: Aug 25 2004 - 13:56
Posts: 1975
Thx Guys!!

Eudimorphodon, you answered my question. Thank you very much. I know that the 867MHz G4 outbeats a 1.7GHz P4, But that is my only reference to go on as fare as mac (oh, and entry level Pentiums and PowerPC's)

tmtomh's picture
Offline
Last seen: 1 year 7 months ago
Joined: Dec 20 2003 - 10:38
Posts: 568
Tempest in a Teacup?

I too am inclined to believe Eudi's claim, based on his experience benchmarking.

But I have to ask - just how much faster is a 300MHz B&W G3 machine than a 9600/350? IIRC, that 350MHz Mach V 604e, with its inline cache, help up pretty well against G3s of similar clock speed.

Now, I know some of that was due to the 604e's butt-kicking FPU, which isn't that important in most everyday use. And I'm also aware that the B&W machines had system busses twice as fast as the 9600's. Nevertheless, I'm inclined to think the overall CPU benchmark speed wouldn't be all that different between the two.

Thoughts?

Matt

Eudimorphodon's picture
Offline
Last seen: 3 months 3 weeks ago
Joined: Dec 21 2003 - 14:14
Posts: 1207
Re: Tempest in a Teacup?

Now, I know some of that was due to the 604e's butt-kicking FPU, which isn't that important in most everyday use. And I'm also aware that the B&W machines had system busses twice as fast as the 9600's. Nevertheless, I'm inclined to think the overall CPU benchmark speed wouldn't be all that different between the two.

I'd agree with you, basically. The 350Mhz 604e did benchmark about the same as the 266Mhz G3s on floating point. However, I'd still lean G3, because the K6-II series *excelled* in integer operations. (Something not reflected in the benchmarks I did.)

That's one reason I'm being nitpicky about saying B&W G3. The other is that the question was "what's the (roughly) equivalent Mac to my system", not "What Mac CPU is equivalent to a K6-II CPU." The memory, video, and bus speeds of the K6 system are substantially superior to a 9x00 series, while they're identical to the G3. And the G3 likewise also has a UDMA/33 IDE hard disk, vs. SCSI on the 9x00. Those items, *plus* CPU performance, seem to give the B&W G3 the best claim for equivalency.

--Peace

Jon
Jon's picture
Offline
Last seen: 12 years 10 months ago
Joined: Dec 20 2003 - 10:38
Posts: 2804
From what I've been recently

From what I've been recently reading about x86 cache designs of that era, don't put more than 128MB of RAM int there, or it could run aweful slow. It will use the RAM abouve the 128MB first, which is outside of the cachable area, so you'll be running as slow as if the cache was turned off. Ack.

Oh, and the board won't support a PII, as that is a P6 design for Slot 1, and won't work in a Socket 7 or a SuperS7. Did you mean P-I (ie 75-166MhZ) and the further MMX versions? That'd make alot more sense. Smile

coius's picture
Offline
Last seen: 10 years 2 weeks ago
Joined: Aug 25 2004 - 13:56
Posts: 1975
I'm Sorry

Oh, and the board won't support a PII, as that is a P6 design for Slot 1

I just realized that. I am thinking about putting a K6-III into it to more of a perfomance boost. Right now I am looking for an MotherBoard that will support a Socket 462 CPU. (I have an Athlon Processor that i am dying to try out) and would trade the Old Motherboard and K6-2 Proc. for just a board that supports the athlon. (I also have other stuff to trade
(see: http://www.applefritter.com/node/6420 ) I will need a board that will have a DRAMM PC100/133 Ram slot. For the athlon

The Czar's picture
Offline
Last seen: 13 years 2 weeks ago
Joined: Dec 20 2003 - 10:38
Posts: 287
Re: Just to know...

To add a bit of weight to my estimation... ;^>

A couple years ago I ran an old C floating-point benchmark on some random machines I had access to.

Here's the results.

<>

Eudi,

Are more details available for those benchmarks you ran? IE: What each of the modules was? I find your results quite intriguing.

Do you remember what OSes were running on the machines at the time? I know in my own experience, when benchmarking the same machine, the OS can make a rather substantial difference in some cases.

Cheers,

The Czar

Eudimorphodon's picture
Offline
Last seen: 3 months 3 weeks ago
Joined: Dec 21 2003 - 14:14
Posts: 1207
Re: Just to know...

Eudi,

Are more details available for those benchmarks you ran? IE: What each of the modules was? I find your results quite intriguing.

Do you remember what OSes were running on the machines at the time? I know in my own experience, when benchmarking the same machine, the OS can make a rather substantial difference in some cases.

I was actually a bit annoyed with myself, because I *know* I had a more current version of that table lying around, and I couldn't find it. (It included results on machines up to 2.8 Xeons and a G4, among other things.) It also had notes on the OSes running at the time.

Anyway, this is the benchmark:

http://home.iae.nl/users/mhx/flops.c

(Also note another flaw in my table. It cut off the four differenty-weighted aggregate "MFLOPS" average at the end. Having wouldn't of changed the order of the table, but it's still annoying I didn't keep it.)

If I recall correctly, the x86 machines were all Linux, (Slackware 8.1 or Red Hat 8.), the iMac was Yellow Dog Linux 2.3, and the iBook was OS X.1. I retested the iMac under OS X shortly thereafter and its score wasn't significantly different. (It averaged slightly lower.)

I am somewhat curious whether the PPC numbers would be better with a newer GCC. YDL and X.1 were both using 2.95 at the time, and allegedly there's been substantial improvement in PPC code generation in the 3.x series. Of course, x86 optimizations have improved as well, so I don't know if there will be any real difference in how the machines compare to each other.

Something to do if it's a boring weekend, I guess.

--Peace

Eudimorphodon's picture
Offline
Last seen: 3 months 3 weeks ago
Joined: Dec 21 2003 - 14:14
Posts: 1207
Re: Just to know...

Something to do if it's a boring weekend, I guess.

It was a boring night. So if anyone's bored:

http://www.applefritter.com/~eudi/flopsbench.html

I've included notes on the hardware and software configurations of the machines, and what compiler switches I used. (I was a lot more aggressive this time in trying to find optimum compiler settings. Not to say that these are the best you could possibly do.)

Not to give away the punch line, but... in at least this synthetic benchmark PowerPC just really isn't that much to write home about. The Sparc CPU did a lot better then I expected, however.

Anyway. I suppose this is slightly off-topic, since I no longer have a K6 machine to test against, so it's missing from the table. ;^>

--Peace

moosemanmoo's picture
Offline
Last seen: 9 years 3 months ago
Joined: Aug 17 2004 - 15:24
Posts: 686
I did some unscientific tests

I did some unscientific tests in 2002 on demo machines at CompUSA, and found a dual 1.25GHz G4 held its own against a 2.8GHz P4 in MIPS, and exceeded by far in MFLOPS. Just for some extra data.

Eudimorphodon's picture
Offline
Last seen: 3 months 3 weeks ago
Joined: Dec 21 2003 - 14:14
Posts: 1207
Re: I did some unscientific tests

I did some unscientific tests in 2002 on demo machines at CompUSA, and found a dual 1.25GHz G4 held its own against a 2.8GHz P4 in MIPS, and exceeded by far in MFLOPS. Just for some extra data.

It's not really "data" unless there's some information on what the test you were running was, obviously.

--Peace

Log in or register to post comments