Apple G5 Benchmark Controversy Continues....

Omega65Omega65 Philadelphia, Pa
edited June 2003 in Science & Tech
from <a href=http://www.digitalvideoediting.com/2003/06_jun/editorials/cw_editorial79.htm>DigitalVideoEditing.com</a&gt;

<i>excerpt</i>

"Nor was the G5 the first personal computer to sport 64-bit capability. The nod here would also go to BOXX Technologies. While talking with Apple and IBM higher-ups, I confronted them with this fact, asking why they were saying the G5 is the "world's first 64-bit desktop" and got an "I'll get back to you" response (read the interview by clicking here). When nobody from Apple ever did get back to me, I pressed the company via email for a comment on this discrepancy. The answer just made me laugh, insulting the intelligence of anybody who might be thinking of buying a product from these snake-oil salesmen. My question was this: "BOXX Technologies has been shipping a dual Opteron 244 1.8GHz machine since June 4. Here's the release we posted when it shipped: <a href=http://www.digitalvideoediting.com/2003/06_jun/news/boxx063.htm.[/url] Does Apple have a response or any comment about that, and does the company still claim to have the first 64-bit desktop computer?" Here's the exact quote from Apple, responding to my question: "The BOXX system you mentioned is marketed as a workstation. There have been other 64-bit servers and workstations available from other vendors that use Itanium, Alpha, and now Opteron. The Power Mac G5 is the first 64-bit personal computer." Oh, so that's what we're doing here. I see -- it's not really truth, it's just marketing. If you have an 8x AGP Pro graphics subsystem, dual processors, every kind of I/O available, fast disks and 8 gigs of RAM, and it's still not a workstation, I don't know what the hell is. All I can say is, if it walks like a duck, quacks like a duck, or looks like a duck, it's probably a duck."

<a href=http://www.digitalvideoediting.com/2003/06_jun/editorials/cw_editorial79.htm>more here</a>

Comments

  • a2jfreaka2jfreak Houston, TX Member
    edited June 2003
    Suprise you?

    When the G4 came out Steve Jobs was spitting forth lies about the G4 pipeline vs P4 pipeline. Apple showed an animation illustrating the problems of a longer pipeline.

    The problem w/ the video that I saw (perhaps there were other videos) is that the pipeline of the P4 kept mispredicting branches, so the pipe would have to be emptied. The G4 obviously has psychic abilities because the branch it took was never wrong, so the pipe always stayed full.

    Now, perhaps there were other videos that corrected this to show a more real-life situation, but I doubt it. Apple doesn't need to tell the truth. Mac zealots will buy Apple because it's Apple, and a few PC guys might buy Apple because of the lies.
    Omega65 said
    from <a href=http://www.digitalvideoediting.com/2003/06_jun/editorials/cw_editorial79.htm>DigitalVideoEditing.com</a&gt;

    <i>excerpt</i>

    "Nor was the G5 the first personal computer to sport 64-bit capability. The nod here would also go to BOXX Technologies. While talking with Apple and IBM higher-ups, I confronted them with this fact, asking why they were saying the G5 is the "world's first 64-bit desktop" and got an "I'll get back to you" response (read the interview by clicking here). When nobody from Apple ever did get back to me, I pressed the company via email for a comment on this discrepancy. The answer just made me laugh, insulting the intelligence of anybody who might be thinking of buying a product from these snake-oil salesmen. My question was this: "BOXX Technologies has been shipping a dual Opteron 244 1.8GHz machine since June 4. Here's the release we posted when it shipped: <a href=http://www.digitalvideoediting.com/2003/06_jun/news/boxx063.htm.[/url] Does Apple have a response or any comment about that, and does the company still claim to have the first 64-bit desktop computer?" Here's the exact quote from Apple, responding to my question: "The BOXX system you mentioned is marketed as a workstation. There have been other 64-bit servers and workstations available from other vendors that use Itanium, Alpha, and now Opteron. The Power Mac G5 is the first 64-bit personal computer." Oh, so that's what we're doing here. I see -- it's not really truth, it's just marketing. If you have an 8x AGP Pro graphics subsystem, dual processors, every kind of I/O available, fast disks and 8 gigs of RAM, and it's still not a workstation, I don't know what the hell is. All I can say is, if it walks like a duck, quacks like a duck, or looks like a duck, it's probably a duck."

    <a href=http://www.digitalvideoediting.com/2003/06_jun/editorials/cw_editorial79.htm>more here</a>
  • ThraxThrax 🐌 Austin, TX Icrontian
    edited June 2003
    As Intel has espoused it's deliriously long pipeline as a fantastic advantage (WOWIES! NETBURST CACHE FO' REAL?!)...Apple has always done the same with their debilitatingly short pipelines.

    Seems AMD's the only one who just pushes it out there and lets the numbers do their thing.
  • a2jfreaka2jfreak Houston, TX Member
    edited June 2003
    Yes, Intel is guilty of acting like NetBurst is the most wonderful thing ever, but AMD's latest QuantiSpeed Ratings are becoming a bit of a joke. the Athlon XP 3200+ should have been a 3000+ or even lower, like 2800+. But still AMD offers an excellent value for the "low-end" CPUs.

    I've puchased w/in the last two months 4 AMD CPUs.
    Thrax said
    As Intel has espoused it's deliriously long pipeline as a fantastic advantage (WOWIES! NETBURST CACHE FO' REAL?!)...Apple has always done the same with their debilitatingly short pipelines.

    Seems AMD's the only one who just pushes it out there and lets the numbers do their thing.
  • Omega65Omega65 Philadelphia, Pa
    edited June 2003
    a2jfreak said
    Yes, Intel is guilty of acting like NetBurst is the most wonderful thing ever, but AMD's latest QuantiSpeed Ratings are becoming a bit of a joke. the Athlon XP 3200+ should have been a 3000+ or even lower, like 2800+. But still AMD offers an excellent value for the "low-end" CPUs.

    I've puchased w/in the last two months 4 AMD CPUs.

    First and Foremost AMD PR rating is based on a Thunderbird class CPU, not a Pentium 4. A Barton 2.20ghz (3200+) = 3200mhz Thunderbird CPU. So the PR may be accurate from a BenchMarketing point of view but not from a Benchmarking point of view (in comparison with the P4).

    There have been 4 or 5 versions of the P4, (P4 256K 400mhz, P4 512K 400, P4 512K 533mhz, & P4 512K 800mhz). Which should the PR refer too...?

    Next the Tbred and Barton rating were accurate in relating to the Pentium 4 until the latest Barton versions.

    Athlon (256K 333mhz) 2.167 = 2700+
    Athlon (256K 333mhz) 2.250ghz = 2800+
    Athlon (512K 333mhz) 2.167ghz = 3000+ (should be 2900+)
    Athlon (512K 400mhz) 2.200ghz = 3200+ (should be 3000+)


    IMO To compare to the P4 (800) a 400mhz 512K Athlon needs 75% of the P4s Clock speed.

    Athlon (512K 400mhz) 2.00ghz = 2700+ (rounded up)
    Athlon (512K 400mhz) 2.10ghz = 2800+
    Athlon (512K 400mhz) 2.20ghz = 2900+ (rounded down)
    Athlon (512K 400mhz) 2.30ghz = 3100+ (rounded up)
    Athlon (512K 400mhz) 2.40ghz = 3200+
  • a2jfreaka2jfreak Houston, TX Member
    edited June 2003
    The QuantiSpeed rating AMD made was said to reflect performance in comparison to a TBird Athlon, but we all know the truth. However, let us assume for a few moments that AMD was being honest and not trying to draw a connection between its QS-rating and the MHz of the p4.

    What benchmarks show an Athlon XP 1.33GHz performing equivalently to an Athlon TBird 1.5GHz? What benchmarks show an Athlon XP 1.4GHz equaling an Athlon TBird 1.6GHz? What benchmarks show an Athlon XP 1.47GHz on-par with an Athlon TBird 1.7GHz?

    Besides, it doesn't matter if the QS-rating is meant as a direct comparison to the performance of the TBird line or the p4 line . . . the Athlon XP 3200+ is just a bad QS-rating! It still gets bested by the 2800+ in raw processing ability and normally doesn't reflect a noticable--definitely not tangible--difference compared to the 2800+ or 3000+.

    As for which p4 should the QS-rating target: none according to AMD. But for most users I think they would say it should be targeted towards the latest p4. Sure the 2400+ won't beat a 2.4GHz p4 w/ an 800MHz FSB, but when the 2400+ came out the p4 running at 2.4GHz was running on a 533MHz FSB. The Athlon XP 3200+ should have been targeted at the p4 3.2GHz which currently is only available with an 800MHz FSB.

    I'll agree the QS-ratings were pretty accurate, if not generally a bit conservative, until the last incarnations, but then the only chip I named was the 3200+ as I really don't have a problem with the 2400+, 2500+, 2600+ or 2800+ . . . though perhaps the 3000+ should have been another 2800+ but what's done is done.

    // Edit:
    I agree with your 75% clock speed idea; I feel it would be a much better match, but unfortunately I don't think we'll see that. Perhaps the Athlon64 will bring a bit more parity to the scene.
    Omega65 said
    a2jfreak said
    Yes, Intel is guilty of acting like NetBurst is the most wonderful thing ever, but AMD's latest QuantiSpeed Ratings are becoming a bit of a joke. the Athlon XP 3200+ should have been a 3000+ or even lower, like 2800+. But still AMD offers an excellent value for the &quot;low-end&quot; CPUs.

    I've puchased w/in the last two months 4 AMD CPUs.

    First and Foremost AMD PR rating is based on a Thunderbird class CPU, not a Pentium 4. A Barton 2.20ghz (3200+) = 3200mhz Thunderbird CPU. So the PR may be accurate from a BenchMarketing point of view but not from a Benchmarking point of view (in comparison with the P4).

    There have been 4 or 5 versions of the P4, (P4 256K 400mhz, P4 512K 400, P4 512K 533mhz, & P4 512K 800mhz). Which should the PR refer too...?

    Next the Tbred and Barton rating were accurate in relating to the Pentium 4 until the latest Barton versions.

    Athlon (256K 333mhz) 2.167 = 2700+
    Athlon (256K 333mhz) 2.250ghz = 2800+
    Athlon (512K 333mhz) 2.167ghz = 3000+ (should be 2900+)
    Athlon (512K 400mhz) 2.200ghz = 3200+ (should be 3000+)


    IMO To compare to the P4 (800) a 400mhz 512K Athlon needs 75% of the P4s Clock speed.

    Athlon (512K 400mhz) 2.00ghz = 2700+ (rounded up)
    Athlon (512K 400mhz) 2.10ghz = 2800+
    Athlon (512K 400mhz) 2.20ghz = 2900+ (rounded down)
    Athlon (512K 400mhz) 2.30ghz = 3100+ (rounded up)
    Athlon (512K 400mhz) 2.40ghz = 3200+
  • Omega65Omega65 Philadelphia, Pa
    edited June 2003
    a2jfreak said
    What benchmarks show an Athlon XP 1.33GHz performing equivalently to an Athlon TBird 1.5GHz? What benchmarks show an Athlon XP 1.4GHz equaling an Athlon TBird 1.6GHz? What benchmarks show an Athlon XP 1.47GHz on-par with an Athlon TBird 1.7GHz?

    As for which p4 should the QS-rating target: none according to AMD. But for most users I think they would say it should be targeted towards the latest p4. Sure the 2400+ won't beat a 2.4GHz p4 w/ an 800MHz FSB, but when the 2400+ came out the p4 running at 2.4GHz was running on a 533MHz FSB. The Athlon XP 3200+ should have been targeted at the p4 3.2GHz which currently is only available with an 800MHz FSB.

    The Benchmarks and audited results are posted on AMD's website.

    When the 2400+ was released the P4 2.4 (400) was what it was compared against. Now there are 3 P4 2.4s (400, 533 & 800) and a Celeron 2.4ghz. Which one is the 2400+ to be compared with?

    On my system I just run the 2400+ & 2500+ at 2.00ghz (FSB400) and don't worry about the rating. :)
  • TekGamerTekGamer Earth
    edited June 2003
    Omega a2 makes a valid point.. One gets the impression that the AMD rating is an equiv. to a p4 mhz, and most people view it that way and refer to it like that.

    On that note, a2 is also right that 3200+ does not perform on par with a p4 3200, its more on par with the 3000.

    Id also go on to say that if I was buying top of the line id choose the intel 3200 as in benchmarks it beat the 3200+ in almost every benchmark. If I am not buying top of the line and want just a regular machine that I AM not gonna overclock, I would choose the Barton 2500+ and If I am going to build a standard overclocker then its P4 2.4c 800 all the way. 1 ghz oc's on air cooling... people are getting free mhz that are = to my duron 1ghz

    At this stage you really cant go wrong with either camp. The speeds are blazing, both are OCin, prices are going down faster then a drunk prom date, its a good time to be a computer parts consumer...

    Tek
  • Omega65Omega65 Philadelphia, Pa
    edited June 2003
    HEY! No middle of the road opinions! :banghead:

    Actually I agree with you.... ;D
Sign In or Register to comment.