Who started saying MHz doesn't matter. I didn't. The fact is, yes MHz does matter. It is also a fact that Apple is correct in stating that there is a MHz myth. My previous comparison of the Alpha to other processors is an excellent example of this. Alpha's rock and they rock hard. A 21362 posts almost 3 times the FP performance of an Athlon in SpecFP.
Now If people would carefully read what I said, I never indicated that a G4 smoked every other processor. [side note, the only 2 machines I currently use at home are a 1.1 Gig K7 and a celleron 466, not a G4... that's my office machine].
What I did state [correctly I believe], was that FAR too much importance is put on MHz. When the G4 is bench'ed with apps that are used in core Apple markets, it's does spank faster processors... even P4s that are twice as fast in MHz. There are a lot of factors behind this. Now is a faster K7 or PIII or P4 going to beat a G4 in non-parallel, interger operations? Probably. They have a higher Ops/Sec than the G4.
don't turn this around and say 'you suck cuz you think MHz doesn't matter', because I've never claimed that. MHz does not matter. I want a dual 800MHz G4 instead of my dual 450 damn it! (actually I want a dual 1.5 GHz G4 but that's even less of an option for me;-).
The point here is that MHz comparisons across different architectures (or even between different chips of the same processor family) can be VERY misleading. If you don't believe me, ask a geek which machine he wants for number crunching... a Joshua, K7, P3, P4, G4, Itanium, Alpha, UltraSparcIII, Power4.... I have a feeling that an educated choice would NOT lead to the highest clocked processor.
What do you think?
Bottom line: G4's SHOULD be faster by now, they ARE faster then significantly higher clocked x86 chips at some task, they ARE slower than x86 processors at other tasks, even at the same clock speed. And, of course, MHz is a bad indicator of overall performance, especially when you look at processors from different architectures... Intel's 'flagship' processor is clocked at less than half of it's middle-of-the-road processor after all.
1)Like I said, there are no new apps out that are not crash test dummy betas. You backed up my point, thanks. Preview releases and promises of future releases don't cut it
Um, software companies can't magically have their products finish development on the same day. It just doesn't work that way. I consider a product annoucement that is within a couple weeks to be a release when it's at an event such as this.
Not wanting to spend too much time, here is what I've found in just about 2 minutes.
iDVD2
Toonboom studio
Suitcase 10
RealBasic 3.5 (august)
Media cleaner 5
Painter 7 (August)
the Sims
Quicken 2002 (later this summer)
Retrospect 5 (fall)
Timbuktu Pro
If they are planning release dates a month or less in the future, they are probably in Final Candidate state... they still need time to press and ship CDs after all.
I'm sure I could find more if I had more time... More vendors will make their announcements over the course of the show anyway.
2) Try to go to the Apple Store, order a new PowerMac G4 and have it delivered this month.
I actually did. The ship date for the 733 model is 2 days! Not too bad if you ask me. That would be this month by the way.;-) To be fair, I did go back and try the 867 and it did list 3-4 weeks ship time. I _believe_ this to be an error since that would push it's ship time back to the same frame as the Dual 800. Jobs made a point to indicate that the low and middle would ship now, and the dual in August.
I wouldn't be suprised if those ship times are based on how long it actually takes to ships a product. Since the 733 has been shipping, it would have an accurate figure (e.g. SAP queries the order database to make an estimate on the last weeks orders and posts that to the Apple Store page...). Since the 867 was just announced, there can't be an accurate ship time since they haven't shipped one yet. I would _guess_ that the ship time on the middle model will change drasticly within a week. Call Apple and ask if you need to know for sure.
3) As much as I would like to think the short pipeline in my Dual 500 G4 makes my machine fast, outside of Photoshop and Media Cleaner Pro, I do not feel like I am driving a HotRod compared to a 1Ghz+ x86 box.
This probably has something to do with the slow Aqua interface (poor finder performance). OSX 10.1 should fix this... this is a software issue, not a hardware issue. The fact that these machines DO routinely trounce much higher clocked x86 machines in A/V benchmarks, and in apache benchmarks (for Darwin running Apache, not for OSX) goes to show that the claims of Apple, though skewed tword Apple's core customers tasks, are valid. (wow, nice run on sentence huh?). You original assertion was that the "MHz doesn't matter" thing was dumb. I still claim that it is an important point, supported by many different platforms (Alpha's are good examples, though I don't consider the G4 an Alpha killer by any means).
I didn't see everything that I wanted but overall, It was a good anouncement.
There are no apps (unless you count the crash test dummy betas)
Um, VPC is almost ready, the preview is downloadable today (very important for me), the new Illustrator is very cool as well as the web integration of Quark totally rocks, WCIII is on shedule for simultaneous release, Word looks good. Even if these don't ship today, at least they are progressed far enough that they look very good in a presentation. I'm pretty happy with the software announcements so far.
On the hardware side, one big YAWN! Speed bumped PowerMac G4s that you can't order and receive for another month. Nothing new in the notebook area and the iMacs are getting pretty lame -- only a RAM boost and a price increase
Well, you are wrong. Plain and simple.
I'm dissapointed that UMA 2 with DDR didn't make it but the high end mac is now the lowest model, and the dual 800MHz model is a huge jump over a single 733MHz at the top of the line.
Not shipping? Um, did you watch the same keynote? 733 and 867 ship today... dual 800's in a couple weeks (august).
Now as for you claim of increased prices (the bit you are truely wrong about), I told the boss to hold off on buying 4 Macs till the conference. For $300 LESS per machine we now get 200MHz faster processor, double the memory, and a CDR/DVD drive instead of a regular DVD. We even get a free inkjet printer with every machine... And this is more expensive in what way?
The really lame part of the hardware demo was Apple's hardware guru telling us, "Mhz doesn't matter"
Why do you bother reading/. ? I thought this site was for tech people.
Since when has MHz been a good indicator of overall performance? What's faster, a 1GHz Athlon or a 1GHz PIII? What's faster, a 1.2 GHz Athlon or a 1.5 GHz P4? (in almost every benchmark the Athlon is faster). Now for different instruction sets, what's faster... a 800 MHz 21364 Alpha, a 1GHz PIII, or a 1.8 GHz Pentium IV? The Alpha's FP performance smokes the other chips. The SpecFP scores for the new Alphas are truely staggering, so much so that Intel feels the need to kill them off within the next couple years, hence the buyout of the Alpha from Compaq.
Is is so hard to believe that the G4, with Altivec and a short pipeline can't outperform a chip clocked twice as fast in A/V and compression benchmarks? Especially when the opponent is routinely smoked by other i386 chips that are 50-60% slower in clock speed? Do you deny that a 20 stage pipeline is little more than a 'trick' to push clock speeds up? Why not crank up the core voltage too and just ship them overclocked from the factory (I think they did this too... or was that AMD... might have been the failed 1.13 GHz P3... someone tried it).
Slashdot Mirror
Now If people would carefully read what I said, I never indicated that a G4 smoked every other processor. [side note, the only 2 machines I currently use at home are a 1.1 Gig K7 and a celleron 466, not a G4... that's my office machine].
What I did state [correctly I believe], was that FAR too much importance is put on MHz. When the G4 is bench'ed with apps that are used in core Apple markets, it's does spank faster processors... even P4s that are twice as fast in MHz. There are a lot of factors behind this. Now is a faster K7 or PIII or P4 going to beat a G4 in non-parallel, interger operations? Probably. They have a higher Ops/Sec than the G4.
don't turn this around and say 'you suck cuz you think MHz doesn't matter', because I've never claimed that. MHz does not matter. I want a dual 800MHz G4 instead of my dual 450 damn it! (actually I want a dual 1.5 GHz G4 but that's even less of an option for me ;-).
The point here is that MHz comparisons across different architectures (or even between different chips of the same processor family) can be VERY misleading. If you don't believe me, ask a geek which machine he wants for number crunching... a Joshua, K7, P3, P4, G4, Itanium, Alpha, UltraSparcIII, Power4.... I have a feeling that an educated choice would NOT lead to the highest clocked processor.
What do you think?
Bottom line: G4's SHOULD be faster by now, they ARE faster then significantly higher clocked x86 chips at some task, they ARE slower than x86 processors at other tasks, even at the same clock speed. And, of course, MHz is a bad indicator of overall performance, especially when you look at processors from different architectures...
Intel's 'flagship' processor is clocked at less than half of it's middle-of-the-road processor after all.
Um, software companies can't magically have their products finish development on the same day. It just doesn't work that way. I consider a product annoucement that is within a couple weeks to be a release when it's at an event such as this.
Not wanting to spend too much time, here is what I've found in just about 2 minutes.
If they are planning release dates a month or less in the future, they are probably in Final Candidate state... they still need time to press and ship CDs after all.
I'm sure I could find more if I had more time... More vendors will make their announcements over the course of the show anyway.
2) Try to go to the Apple Store, order a new PowerMac G4 and have it delivered this month. ;-) To be fair, I did go back and try the 867 and it did list 3-4 weeks ship time. I _believe_ this to be an error since that would push it's ship time back to the same frame as the Dual 800. Jobs made a point to indicate that the low and middle would ship now, and the dual in August.
I actually did. The ship date for the 733 model is 2 days! Not too bad if you ask me. That would be this month by the way.
I wouldn't be suprised if those ship times are based on how long it actually takes to ships a product. Since the 733 has been shipping, it would have an accurate figure (e.g. SAP queries the order database to make an estimate on the last weeks orders and posts that to the Apple Store page...). Since the 867 was just announced, there can't be an accurate ship time since they haven't shipped one yet. I would _guess_ that the ship time on the middle model will change drasticly within a week. Call Apple and ask if you need to know for sure.
3) As much as I would like to think the short pipeline in my Dual 500 G4 makes my machine fast, outside of Photoshop and Media Cleaner Pro, I do not feel like I am driving a HotRod compared to a 1Ghz+ x86 box.
This probably has something to do with the slow Aqua interface (poor finder performance). OSX 10.1 should fix this... this is a software issue, not a hardware issue. The fact that these machines DO routinely trounce much higher clocked x86 machines in A/V benchmarks, and in apache benchmarks (for Darwin running Apache, not for OSX) goes to show that the claims of Apple, though skewed tword Apple's core customers tasks, are valid. (wow, nice run on sentence huh?). You original assertion was that the "MHz doesn't matter" thing was dumb. I still claim that it is an important point, supported by many different platforms (Alpha's are good examples, though I don't consider the G4 an Alpha killer by any means).
There are no apps (unless you count the crash test dummy betas)
Um, VPC is almost ready, the preview is downloadable today (very important for me), the new Illustrator is very cool as well as the web integration of Quark totally rocks, WCIII is on shedule for simultaneous release, Word looks good. Even if these don't ship today, at least they are progressed far enough that they look very good in a presentation. I'm pretty happy with the software announcements so far.
On the hardware side, one big YAWN! Speed bumped PowerMac G4s that you can't order and receive for another month. Nothing new in the notebook area and the iMacs are getting pretty lame -- only a RAM boost and a price increase
Well, you are wrong. Plain and simple.
I'm dissapointed that UMA 2 with DDR didn't make it but the high end mac is now the lowest model, and the dual 800MHz model is a huge jump over a single 733MHz at the top of the line.
Not shipping? Um, did you watch the same keynote? 733 and 867 ship today... dual 800's in a couple weeks (august).
Now as for you claim of increased prices (the bit you are truely wrong about), I told the boss to hold off on buying 4 Macs till the conference. For $300 LESS per machine we now get 200MHz faster processor, double the memory, and a CDR/DVD drive instead of a regular DVD. We even get a free inkjet printer with every machine... And this is more expensive in what way?
The really lame part of the hardware demo was Apple's hardware guru telling us, "Mhz doesn't matter" /. ? I thought this site was for tech people.
Why do you bother reading
Since when has MHz been a good indicator of overall performance? What's faster, a 1GHz Athlon or a 1GHz PIII? What's faster, a 1.2 GHz Athlon or a 1.5 GHz P4? (in almost every benchmark the Athlon is faster). Now for different instruction sets, what's faster... a 800 MHz 21364 Alpha, a 1GHz PIII, or a 1.8 GHz Pentium IV? The Alpha's FP performance smokes the other chips. The SpecFP scores for the new Alphas are truely staggering, so much so that Intel feels the need to kill them off within the next couple years, hence the buyout of the Alpha from Compaq.
Is is so hard to believe that the G4, with Altivec and a short pipeline can't outperform a chip clocked twice as fast in A/V and compression benchmarks? Especially when the opponent is routinely smoked by other i386 chips that are 50-60% slower in clock speed? Do you deny that a 20 stage pipeline is little more than a 'trick' to push clock speeds up? Why not crank up the core voltage too and just ship them overclocked from the factory (I think they did this too... or was that AMD... might have been the failed 1.13 GHz P3... someone tried it).