1 GHz G4 Processor = ?

verces

macrumors member
Original poster
Oct 13, 2004
44
0
What exactly is the equivalent on PC terms? 1 Gigahertz of Mac G4 is equal to? Intel or AMD Processing speed of?
 

RacerX

macrumors 65832
Aug 2, 2004
1,504
2
Well, I'm not sure there is really any way to compare a Windows PC with a Mac running Mac OS X. I, personally, can't work with a system as crippled as Windows.

But, removing Windows from the equation here and looking at raw performance... the best study I've seen was a performance comparison done by a NASA engineer.

His findings were ruffly (in terms of MFLOP/MHz):
  • Pentium 4: 0.096
  • G4: 0.105
  • G5: 0.127
The Pentium was running Red Hat.

You asked about a G4 1 GHz specifically, so then these processors at these speeds would be about equal:
  • G4/1000 MHz
  • Pentium 4/1094 MHz
  • G5/827 MHz
assuming that the performance is linear across all factors, which it isn't as things like bus speed and cache size can effect these numbers greatly.

Hope this helps out a little.
 

johnnyjibbs

macrumors 68030
Sep 18, 2003
2,958
119
London, UK
I have a 1GHz G4 PowerBook (133MHz FSB) and a 2.4GHz Pentium 4 (533MHz FSB) Dell desktop and they are very evenly matched, regarding Unreal Tournament 2004, Photoshop Elements and general useage.

EDIT: The numbers in the post quoted above don't look right to me. The G4 is much more efficient than the Pentium 4 (7 stage pipeline compared with 20) and is more efficient than even the G5. This means that a G4 should outperform a G5 slightly at the same clockspeed, and significantly outperform the Pentium 4 at the same clockspeed.

I would say that generally..

1 GHz G4 is roughly equal to a 1.3 GHz G5 and a 2.4 GHz Pentium 4.

In other words, don't worry about Macs looking underpowered.
 

paxtonandrew

macrumors 6502
May 10, 2004
323
0
I Come From A Land Down Under
I have a 667 G4 that runs rings around my 1 gig Pentium M, but the computer is 9 months old, and running Xp Service Pack 2, which is acting as such a drain on the CPU that it is shocking. My PowerBook 1.33 makes it look like a primitive piece of crap, but again, that is the operating system. Windows needs a re-code.
 

vasaz

macrumors member
Mar 23, 2004
51
0
MN
RacerX said:
His findings were ruffly (in terms of MFLOP/MHz):
  • Pentium 4: 0.096
  • G4: 0.105
  • G5: 0.127
I don't dispute the above scores, but keep in mind the MFLOPS = millions of floating operations per second. So it's a measure of mostly only floating point stuff, not a general system performance.
 

RacerX

macrumors 65832
Aug 2, 2004
1,504
2
johnnyjibbs said:
...The numbers in the post quoted above don't look right to me. The G4 is much more efficient than the Pentium 4 (7 stage pipeline compared with 20) and is more efficient than even the G5. This means that a G4 should outperform a G5 slightly at the same clockspeed, and significantly outperform the Pentium 4 at the same clockspeed.

I would say that generally..

1 GHz G4 is roughly equal to a 1.3 GHz G5 and a 2.4 GHz Pentium 4.
Well, Apple considered the single processor G5 at 1.8 GHz to be ruffly equal to the previous top of the line PowerMac, the Dual G4 at 1.42 GHz

The G5 is faster than a G4 of the same clock speed.

As for the rest of the scores, this was looking at code that could be executed on all three processors. When Altivec processing was added in, the scores jumped up dramatically, but the Pentium 4 system was left out of the benchmarking.

Further, all benchmarks were single processor. Most PowerMacs in recent years have been released as dual processor systems, while the Pentium 4 can not be used in a dual processor system and requires a special license to run Windows XP on dual processor systems (like the Xeons which can run in multi processor systems).

When the scores were done where both processors on the G4 and G5 systems were in use, the scores (predictably) just about doubled.

So a dual G4/1 GHz would be ruffly the same as a Pentium 4/2.2 GHz, or a single processor G5/1.6 GHz.

Also, yes, these are floating point scores. As most real work on computers use floating point, that is, in my opinion, the only score that matters when comparing processors. Integer performance is nice for things like a work processor, but I highly doubt that the fastest integer system in the world would make you a faster typist.

I don't use Windows PCs. I guess they are nice for gaming and secretaries, but I don't see much value in them beyond that. My background is in Macs, Suns and SGIs, and the main thing those companies worried about was floating point performance.

And as I pointed out, Windows was not in use in this study. The G4s were running Mac OS X 10.2.3, the G5s were running Mac OS X 10.2.7 and the Pentium 4 systems were running Red Hat Linux 7.3. I have no doubt that a Windows PC would run slower, but this study didn't include the crippling effect of Windows on a system.

So, anyone can dispute these figures if they want, but this was the best study I've seen that attempted to bench mark these processors on a level playing field. I, personally, would have benched the systems using the altivec as enough software takes advantage of it to include a head to head comparison. But that was not how it was done, and I'm not into cooking the books when reporting the findings on someone else's work.
 

johnnyjibbs

macrumors 68030
Sep 18, 2003
2,958
119
London, UK
RacerX said:
Well, Apple considered the single processor G5 at 1.8 GHz to be ruffly equal to the previous top of the line PowerMac, the Dual G4 at 1.42 GHz

The G5 is faster than a G4 of the same clock speed.
If a 1.8GHz G5 is roughly equivalent to a [dual] 1.42GHz G4, then a G5 at the equivalent clockspeed (1.42GHz) would be slower than the G4, not faster, as you stated. I don't believe that dual processors ups the speed that much, only in certain tasks.

I'm not doubting that Windows would slow it down or that Macs are that slow, but I think the G4 is more powerful than you are giving credit.

Oh, and by the way, "ruffly" is spelt "roughly" ;)
 

RacerX

macrumors 65832
Aug 2, 2004
1,504
2
johnnyjibbs said:
If a 1.8GHz G5 is roughly equivalent to a [dual] 1.42GHz G4, then a G5 at the equivalent clockspeed (1.42GHz) would be slower than the G4, not faster, as you stated. I don't believe that dual processors ups the speed that much, only in certain tasks.
Yeah, the tasks that Apple was benchmarking, Photoshop, Final Cut Pro, and other high end applications that the PowerMac was design for.

But sorry, the G5 is faster than the G4 at the same clock speed.

Oh, and by the way, "ruffly" is spelt "roughly" ;)
Thanks. Ruffly was a rough spelling of roughly.

Fortunately it doesn't effect the facts at hand. Wishful thinking is hardly a persuasive argument.

Better luck next time. ;)
 

gekko513

macrumors 603
Oct 16, 2003
6,302
1
A 1GHz G4 certainly is better than a 1.094GHz P4. Maybe on some floating point problems that will be the case, but on Altivec floating point, the G4 is both theoretically and in practise better than a 2GHz P4. Source The P4 can't even do normal flops while doing vector flops. On unpredictable integer ops, the G4 is also much better per GHz than a P4. I have gotten results where a 1GHz G4 performs almost like a 1.8GHz P4.

So it's impossible to say anything other than that a 1GHz G4 performs at least as well as a 1.1GHz P4 but not generally better than a 2.5GHz P4 and most of the time somewhere in between.
 

vraxtus

macrumors 65816
Aug 4, 2004
1,043
0
San Francisco, CA
johnnyjibbs said:
I have a 1GHz G4 PowerBook (133MHz FSB) and a 2.4GHz Pentium 4 (533MHz FSB) Dell desktop and they are very evenly matched, regarding Unreal Tournament 2004, Photoshop Elements and general useage.
I have no idea where you're getting your numbers from.

Barefeats showed a *while* ago that Athlons and Pentiums far outperform a G4 on performance apps like UT2K4. Compared to a G5 an Athlon showed more than TWICE the FPS rates for normal botmatches.

As for your Mhz comparisons I really don't think those are on the dot either. There's really no good comparison at the moment between G4 and G5 raw power since the most modern G4s and G5s are duals. IMO the best comparison would be to do a raw performance scalar on a slower SP G4 to a SP G5 and go from there.
 

freiheit

macrumors 6502a
Jul 20, 2004
643
90
California
Perception is the best benchmark

I've not used a single processor Mac -- and I've not used a dual processor PC.

But in my short time using a dual processor 1.25GHz G4 Mac, it performs at least as well, if not better than, my AMD Athlon 2400+ (2GHz) which itself has been shown to perform on average like a 2.4GHz Pentium 4.

There really is no substitute for dual processors :) unless you simply cannot afford it. Even if my most demanding single application doesn't make use of both processors, the OS does and so can juggle two, three, five or a dozen concurrent programs much easier than a single processor system (such as my "faster" PC).

I still use my PC for most things -- I'm transitioning from PC to Mac -- but the Mac definitely feels faster and more responsive. That cannot be quantitatively benchmarked, but IMHO it's the best benchmark. Perception is 90% of reality.
 

J.Allen

macrumors member
Oct 12, 2004
34
0
Adelaide, South Australia
AMD 64/Opteron = God
Everything else = playing catchup:p.

Apple OSX has alot of what I refer to as "Wank factor" eg the onscreen animations, the shocking code that makes up iphoto and itunes among other things that hamper its performance. XP has less of the "Wank Factor" and hence appears to run faster on a system with same clock speed, equal RAM/Grahics card, etc...

$ for $ your going to get better "performance" out of a AMD/intel (Windows) based system than you are out of a IBM/Moto (apple/OSX) based system at this time.
This is a hard fact that can't be debated.
 

QCassidy352

macrumors G4
Mar 20, 2003
10,571
2,661
Bay Area
J.Allen said:
$ for $ your going to get better "performance" out of a AMD/intel (Windows) based system than you are out of a IBM/Moto (apple/OSX) based system at this time.
This is a hard fact that can't be debated.
if it's a hard fact, could we see some numbers and sources?
 

Diatribe

macrumors 601
Jan 8, 2004
4,226
21
Back in the motherland
vraxtus said:
I have no idea where you're getting your numbers from.

Barefeats showed a *while* ago that Athlons and Pentiums far outperform a G4 on performance apps like UT2K4. Compared to a G5 an Athlon showed more than TWICE the FPS rates for normal botmatches.
You cannot really compare a mac to a pc with UT2K4....
If you'd use a game than runs on openGL on a pc then you can compare. Those games are directx optimized and have a far higher optimization than for opengl on a mac. Other tests, using real world apps show different results.
 

jiggie2g

macrumors 6502
Apr 12, 2003
491
0
Brooklyn,NY
Not to start a flame war but i would say after testing many systems myself that it goes like this.

1ghz G4 = 1.8-2.0ghz P4 Northwood
2.0-2.2ghz P4 Prescott

1ghz G4 = Athlon XP +1800
Sempron +2200
(not even gonna go there with Athlon 64/FX)

1.8Ghz G5 = Athlon 64 +2800 (1.8ghz newcastle)

I will say that the Athlon FX chips are much faster than G5's at any speed especially now that the FX-55 1MB L2 cache is out , even Athlon 64's are runing at +4000. I would just get a Run Mandrake 10.1 64 bit (primary OS)/ Win XP Pro(games) Dual Boot on a AMD machine.

Apple and IBM better get on the ball cuz AMD is coming hard and they are going to get left behind like the old G4 day all over again.

AMD has the new Winchester core 939 pin 90nm fab(very overclockable a few sites have OC'ed an Athlon 64 +3000 (1.8ghz) winchester to 2.6 on air(2.4 w/stock fan) and they will be getting PCI-Express Mb's in the form of nforce 4 and Via K8T890 with in the next week.

plus they are already showing off the Dual Core CPU's for mid 2005, so far AMD has come through on every promise they made(can't say the same for IBM).
 

johnnyjibbs

macrumors 68030
Sep 18, 2003
2,958
119
London, UK
vraxtus said:
I have no idea where you're getting your numbers from.

Barefeats showed a *while* ago that Athlons and Pentiums far outperform a G4 on performance apps like UT2K4. Compared to a G5 an Athlon showed more than TWICE the FPS rates for normal botmatches.

As for your Mhz comparisons I really don't think those are on the dot either. There's really no good comparison at the moment between G4 and G5 raw power since the most modern G4s and G5s are duals. IMO the best comparison would be to do a raw performance scalar on a slower SP G4 to a SP G5 and go from there.
I'm going on perception generally, although Photoshop Elements filters take about the same time to complete on my 1GHz PB G4 as my 2.4GHz P4 Dell (OK, so there could be efficiency differences in the coding of the programs but we're talking real world stuff here. I used the same photo and the same filters to make a fair comparison).

Unreal Tournament 2004 (demo) on my Mac runs at a slightly higher framerate than the demo on my PC in the same match (although to be fair, my PB has 5200 Go graphics 32MB VRAM and the Dell has integrated graphics with shared RAM up to 64MB).

But my Mac is far more responsive when it comes general stuff like using the Finder (it takes ages for Windows Explorer windows to launch, and it is a lot less stable). I'm not going to say Word v.X is faster than Word 97 running on my PC but that's another matter ;)

Both systems have 768MB RAM.
 

jiggie2g

macrumors 6502
Apr 12, 2003
491
0
Brooklyn,NY
J.Allen said:
AMD 64/Opteron = God
Everything else = playing catchup:p.

Apple OSX has alot of what I refer to as "Wank factor" eg the onscreen animations, the shocking code that makes up iphoto and itunes among other things that hamper its performance. XP has less of the "Wank Factor" and hence appears to run faster on a system with same clock speed, equal RAM/Grahics card, etc...

$ for $ your going to get better "performance" out of a AMD/intel (Windows) based system than you are out of a IBM/Moto (apple/OSX) based system at this time.
This is a hard fact that can't be debated.

I agree 100% ,while the mac zealots will never admit it, the AMD 64/FX series own's the G5. no one will here acknowledge the fact that Apple still needs 2 G5's just to match the Beanchmark scores of AMD Chips, because in Single vs Single CPU test Apple get's hammered all the time. every test i've seen on Dual Opteron vs Dual G5 , the G5 gets spanked badly.

while I myself own a mac , I love the OS but , most people who just play games , rip & burn music , surf the web, do a little MS office work ...which means about 90% of all users dont really care for Mac , simply cuz they don't need it. a PC does all thst simple stuff just fine. that said however more advanced users can benefit greatly from a mac( if they can afford one).
 

J.Allen

macrumors member
Oct 12, 2004
34
0
Adelaide, South Australia
QCassidy352 said:
if it's a hard fact, could we see some numbers and sources?
www.pricewatch.com
you should be able to get all the prices you need off there.

If not, I'll gladly benchmark my ageing Athlon 1GHz based system against any G4 based computer. (Note this computer is approaching 3.5 years in age, but I think it will give the G4's a run for their money ;))
 

gekko513

macrumors 603
Oct 16, 2003
6,302
1
You're forgetting the Big Mac that totally crushed all competition when it came to price/performance ratio. Including Opteron, Xeon, Itanium and all the others.

It'll be interesting to see the new top 500 list. Maybe some of the competitors are starting to catch up?
 

gekko513

macrumors 603
Oct 16, 2003
6,302
1
And what's this? The dual G5@2GHz beats the dual Opteron@2GHz on Bryce and After Effects but loses on Photoshop. I wouldn't call that "owning".
 

SilentPanda

Moderator emeritus
Oct 8, 2002
9,808
28
The Bamboo Forest
A 1 ghz G4 has roughtly the same processing speed as a a jack-o-lantern.

And you'll never figure it out for real because everybody will disagree about something.
 

verces

macrumors member
Original poster
Oct 13, 2004
44
0
So, all in all...from what I've gathered so far....

Processing speed wise overall, in technology, considering all aspects and basing on very-equal configurations.

1. G5
2. Opteron
3. Pentium 4

and a 1GHz G4 is equivalent to a about 2GHz processing speed of a AMD or an Intel.

Right?
 

johnnyjibbs

macrumors 68030
Sep 18, 2003
2,958
119
London, UK
You're probably about right. Except I think a dual 2.5GHz G5 would beat a 3.6GHz P4 system in most tests (except gaming maybe :rolleyes: )

Do not be alarmed that a dual 2.5 G5 has 2.5 times the clockspeed of a 1GHz G4 - the difference is less, except in very big tasks (video rendering, 3D, etc). A G4 1GHz (what I work with) holds its own favourably and is great for consumers and general work.

It all depends what you want the system for and what to do... All I will say is that a 1GHz G4 is MUCH better than a 1GHz PIII or P4, if that's what you're worried about.