POWERLOGIX COMMENTARY ON NEW G5 ANNOUNCEMENT FROM APPLE

AUSTIN --June 25, 2003 -- PowerLogix congratulates Apple on the announcement of the IBM 970-based PowerMac G5. As an aftermarket supplier of Apple hardware, it is in PowerLogix' best interest to see a large installed base of Apple machines. We want to see Apple succeed in a big way with G5. At the same time, we want to make sure customers have accurate information so they can evaluate what is best for their individual needs...whether to keep their existing G4 and upgrade with a CPU card from PowerLogix; or buy a new G5 and hand down their previous PowerMac G4 to someone else in the office...perhaps installing a CPU upgrade in it at the same time.

To properly evaluate the impact of these new machines, one has to look at several different factors.

In a nutshell, it's all about price vs. performance. If your business needs the absolute ultimate performance, regardless of cost (say, even just 10% faster for an extra couple of thousand dollars), it is probably best to buy a PowerMac G5. If your clientele and business allows you to pay for a G5 in a short time due to the increased productivity, then perhaps it is a no brainer to buy a new machine.

However, it must be said that the majority of computer users (PCs as well as Macs) do not truly 'need' to pay top dollar for the ultimate performance. (Need and want are two different things...if you want ultimate performance and can pay for it, then this discussion doesn't apply.) For this largest group of users, especially in this struggling economy, it's a matter of getting the most bang for the buck. If a user can upgrade a machine for a few hundred dollars and get the lion's share of the performance of a new Mac (which otherwise would cost several thousand dollars) then that is often the best business decision.

It's difficult to do a truly fair performance comparison at this point, because the G5s are not available for independent testing. All we have to work with (as of now) are the published Apple benchmarks from Steve's presentation. We have to take Apple's marketing with a little bit of salt, as, after all, they want (and need) to put the best possible spin on the performance of these new machines. (Refer to this link or this one for two viewpoints regarding Apple's controversial marketing hyperbole.)

We do have dollars and other factors with which to compare, however. The Dual 2GHz G5 is $2999, with 512Mb of RAM.If you are a power user and are using 2.0Gb of RAM in your current machine, you will not be able to take your RAM that is now in your G4 and put it into your G5. This means if you want to have 2.0Gb in your G5 you must to pony up another several hundred dollars for that extra gig and a half. On the Apple store, that will set you back another $750. It's cheaper from third parties, of course, but it's still extra expense that one doesn't incur when upgrading the CPU with a PowerLogix card.

Then there is the hard drive and drive bay issue. There are only two drive bays in the G5 (one CD bay) and the drives use serial ATA. You won't be able to take your drives with you, unless you use up one of your slots with an ATA card, and since there are fewer drive bays, you might have to put that drive in an external enclosure if you run out of physical space (you could use a firewire enclosure and save a slot, however.) In any case, you can't just move your hard drive from inside your G4 to inside your G5. So there is some inconvenience at least, and extra expense in the worst case scenario. Secondarily, additional serial ATA drives cost more, too, although they are faster.

One more thing: the new Macs will not boot into OS9. This is an important issue for many users.

Lastly, a little about performance...as much as we know today. As we pointed out in our DDR white paper, it's not just a question of raw MHz comparisons. We showed that DDR L3 was in the most extreme, best case scenario, only 15% faster than SDR cache memory...even though on a MHz vs MHz basis DDR sounds like it is 2x faster. The same goes for DDR memory; the bus speed of 133MHz on your current G4 might seem very slow compared to the "up to 1GHz bus" on the G5s, but it's definitely NOT anything close to 7.5x faster. Lastly, the pipelines of the G5 are different than pipelines on a G4 7455...so again, one can't exactly compare clock speed to clock speed. (They are different enough that Apple posted this page comparing the two chips and how one should write code for the G5. Note the following technical points from that page: the G5 has direct-mapped cache; this is significantly inferior to the G4 8-way set associative cache. In addition, look at the latency. 135 vs only 93 on the G4. That is a lot of clock cycles...and requires careful coding to avoid performance hits, as the paper outlines. Pipeline lengths: 23 stages for G5, vs 7 stages on G4...this a major contributor to the 1.45x increase in latency: any time the instruction queue has to be emptied due to a branch, or jump, it has to flush and re-fill 3.5 times more data...G5 definitely needs that FSB bandwidth!) Bottom line: it's going to come down to real world applications, and we'll find out soon enough just how fast these machines really are. Once this is known, only then can the true price performance graph be drawn.

By the way, anyone remember when the PowerMac G4 733 with the 7450 CPU first began shipping? Apple touted it as the best thing since...well....the PowerMac G4 with PPC 7400...but the previous 533 model was actually a better performer than the 733! This was primarily due to the pipeline architecture of the 7450 compared to the 7400. Apple needed to get the clock speed up for marketing purposes (then and now.) Will history repeat itself? How close will the previous Dual 1.42 PowerMac G4 come to the new 1.6GHz or 1.8GHz G5s, for example?

Interestingly, using Apple's own Skidmarks GT benchmark application, our Dual 1.4G CPU card in a 100MHz bus early PowerMac G4 AGP (now four years old) scored about 82% of a new G5 on the integer benchmark (a score of 172 for the G5 and 140 for the G4.) (Refer to this link for more information. Please be aware that the G5 was running a new version of OS X which could account for some difference as well.) However 82% of the performance for about 28% of the actual out of pocket cost is a bargain by anyone's measure. This is only one benchmark, of course...so one shouldn't draw comparisons until the G5s actually are shipping.

Hopefully real G5 performance data will be available soon.

One thing is for sure: no matter how fast the G5s are, not everyone needs one. For example, if you have a graphic design firm with 10 G4s, you might find two users who TRULY require maximum speed; those users could opt for new G5s. The other 8 might be completely satisified and productive upgrading their 500 or 533MHz G4s to a dual 1.4Ghz PowerForce card for only $999 each. Net out of pocket: about $15K (as opposed to perhaps $35K for 10 new dual G5s.) (Or they could save even more money buy upgrading some machines to single 1GHz, single 1.4, or perhaps dual 1GHz or dual 1.2.)

PowerLogix remains committed to providing Mac owners with the most economical performance solutions, and we continue to be well-positioned as the foremost manufacturer of the fastest Mac upgrades.

When deciding on buying an upgrade, the obvious choice is PowerLogix. Our large selection of the market's top-performing upgrade cards, and the exclusive supply of Dual G4 upgrades, assures that you are making the right decision.

 

About PowerLogix
PowerLogix manufactures innovative products like the PowerForce, iForce and BlueChip processor upgrades for maximum performance from second generation and newer Apple Power Macs and clones. For more information, contact PowerLogix, 8701 W. Parmer Lane, Suite 1120, Austin, Texas 78729, 512-401-9902, Fax: 512-401-9943, info@powerlogix.com, www.powerlogix.com.

-30-