In the past two installments of this review, covering G4 processor upgrades from Sonnet, PowerLogix and GigaDesign made for the G4 Tower machines, we have gone over the performance and pluses & minuses of the low-end, and middle of the road options these companies offer. The previously reviewed cards run at 800 MHz (at the low-end) and 1 GHz (the middle of the road models). This time we are going to examine the top-of the line models in this family of upgrades. These cards range from 1.2 GHz to 1.25 GHz and have 2 MB of L3 cache. Though they come with a price premium, the cost may be worth it for some upgraders, and perhaps necessary for others.
In our estimation, there are three groups of people that usually consider upgrading their machines: the diehard speed freaks, the high end production workers, and the middle of the road 'meat an tater's' crowd. What all three groups have in common, is that they are looking to get a speed fix on a budget, and are not ready to let the relationship to their current machine go ... even if that relationship is on the cusp of growing stale.
The speed freaks are ... just that. They want every scrap of power they can lay their hands on... without actually shelling out for a new machine. In practical terms they may not actually need all that speed. But this isn't about need .. it is about desire. If it is there ... they want it. And who would deny them their indulgence ... life is too short. These guys (mainly) and gals (some), are addicted to speed. Though somewhere in the back of their brains is a small voice telling them it would make more sense to get the entry level model of the upgrade, it can't be heard over the cerebral cortex, drunk on dopamine, screaming speed, speed ... SPEED! This group is a built in customer base for upgrade manufacturers .. because the current fix doesn't last long. It will only be a short while before they are on the prowl again, impatient for their Pushers to put the next speed bumped processor upgrade on the market. Your humble author used to be afflicted with speedfeakism, but hasn't done a dose of GHz since he joined UA (Upgrader's Anonymous). I have to admit, though I have been sober for a couple of year, my palms sure got sweaty when it came time to return these upgrades (think Frodo & Bilbo at Rivendell, and the ring tug of war). For speed freaks, the high-end upgrade is the natural choice .. why choose one or two scoops when you can have three.
The second group of potential upgraders, for whom the high-end upgrade might be appropriate, and perhaps a necessity, are the production people. They do real work with their computers. These are the people who use Photoshop in a production environment, Final Cut Pro users and anyone who earns their bread running processor intensive applications. For them time is money and a top of the line processor upgrade, even if they have to pay top dollar for it, could pay for itself quickly in time saved. Upgrading processors on a regular schedule will help users in these environments extend the life of their current machines, keeping cost down and avoiding the disruption that inevitably occurs whenever you upgrade to a completely new machine. In most cases you can slap a processor upgrade into a machine and be back on your feet within minutes. The most important commodity, for those in this professional group, is raw processing power. Upgrading other aspects of a machine, drives, graphics card, is usually less important to workflow. If it isn't then these performance categories have probably already been enhanced to their fullest. Adding a processor upgrade into the mix will only continue the computer's ability to greedily consume the rivers of data professionals throw at it. The only caveat for those in this category is, that one processor may not be enough. It is possible to upgrade your aging G4 Tower to a dual processor. PowerLogix makes such a solution. But you need to assess if the applications you use, or your style of workflow, will harness all the extra power of a dual processor setup.
Our middle of the road user engages in the meat and potato processing that most general use computers do. These computers can be found in schools, small businesses, and at home, where they do yeomen work, but usually never engage in major campaigns. These Macintoshes do not need a high-end card. If you want one ... then by all means go out and get one. But I think most of you will see a high-end card like a holiday fling with a high flyer .. briefly exciting, but ultimately not a good match .. Not really what you were looking for. A high-end card is overkill, and you will pay through the nose for it .. without reaping the benefits. You need something, that will; goose OS X performance on your middle aged machine, make some of those more demanding applications flow a little better, and, yes when you need some extra processing power, your sensible machine will respond. Previously in this review, we have suggested that these users opt for the low-end 800 MHz card in this category, eschewing the middle 1 GHz, and the 1 GHz (plus) models listed here. You won't get as much octane ..but then you don't really need it. Gentlemen start your ... Volvos.
Performance
On processor intensive tasks the 1.2 GHz upgrades turned in a 15 to 25% performance improvement over their 1 GHz siblings. On tasks that involved more than straight processor work the amount of performance improvement was somewhat less. Also, as we noted in our reviews of the 800 MHz cards and the 1GHz cards, in drive performance, and also when encoding MP3s using iTunes, there continued to be a slight fall off in performance, over the performance of the Tower machine in stock configuration. A puzzling phenomenon.
On the value front, if you opt for the 1.2 GHz model over the 1 GHz one, you are getting 20% better processor performance, for about a 8% - 24% increase in price.(depending on which manufacture you choose)
In contrast, when compared to the 800 MHz upgrades, the 1.2 GHz cards will give you about a 50% raw processing performance improvement for about a 49% - 90% price premium (depending on which manufacture you choose).
For more depth on the performance of this class of upgrades please take a look at the first and second parts of this review
Stability
Again these cards performed well. We had zero stability problems during our performance and stress testing. It was suggested to us, after our first installment of this review, that because the GigaDesign processor upgrade comes with a specially designed heatsink, with its own dedicated fan, that its cooling system is superior to that of Sonnet or PowerLogix, which makes use of the stock heatsink and cooling system. This may be so, but none of the upgrades showed any signs of heat stress during our testing.
Though PowerLogix has moved to a new unified upgrade for the G4 Towers, the PowerForce G4 Series 233, at the time of our testing they were still working out some bugs with the 1.2 GHz version of the card. Consequently we used a Series 100 card (only works on a G4 machine with a 100 MHz bus ... a Series 133 model works on machines with a 133 MHz bus). However, in terms of performance, the Series 233 cards and the Series 100 cards should perform similarly, at the same given clock speeds. [Editor's note: Update - since this was written PowerLogix has decided to continue with the production of the Series 100 and Series 133 cards. Both the 100 and 133 cards come with their own heatsinks and individual fans]
Conclusion
The three cards performed almost identically. The GigaDesign G-celerator is clocked 50 MHz faster than the others, and so turned in an ever so slightly higher score in raw processing tasks. The the G-celerator also has the lowest price, coming in $50 cheaper than the PowerLogix card and a whopping $131 lower than the Sonnet Card. Our scoring differences below reflect the differential in pricing, more than one cards technical/performance merits over the others. Some may argue that GigaDesign and PowerLogix have superior cooling capabilities, but if this is an issue with the Sonnet card it did not crop up during our testing. It should be noted that price cuts occur on a regular basis with all manufactures trying to stay ahead of the others in offering the best value. It may be that in the next round of price slashing that cost difference between the Sonnet card and the other two will narrow somewhat.
These speedy cards are most appropriate for those with either the need or desire for the most performance possible. It should be noted that PowerLogix also makes dual processor versions of these cards (currently the only one to do so). Again, those after the ultimate in speed, may find a dual processor card more appropriate. You really need to look at the applications you use, and the manner in which you use them Determine if a dual processor setup makes sense given your workflow. While the single processor 1.2 GHz card from PowerLogix will set you back $449, the dual processor version is $799. Not bad if you are going to use all that horsepower, but really over-kill in most situations, outside of a production environment.
If you own a Cube and want a processor upgrade for it, Sonnet & PowerLogix are currently your only options.
[ Update: Since this was written PowerLogix and GigaDesign have announced single processor upgrades in this family running at 1.4 GHz. GigaDesign should be shipping these upgrades by the time you read this ($575), and PowerLogix says there's should be available in a few week's time ($599). According to PowerLogix the 800 MHz and 1 GHz versions of their card will be phased out soon in favor of the faster upgrades]
Hits: Good at processor
intensive tasks, speedy processor caches. Works in Cube. Very good stability. Three year warranty
Misses: More expensive than competition. Installation slightly more involved
Requirements: Power Mac G4 (AGP Graphics), Power Mac G4 (Gigabit Ethernet), Power Mac G4 (Digital Audio), Power Mac G4 (QuickSilver 2001), Macintosh Server G4 (450 & 500 MHz), Cube. OS 9.2 through Mac OS X. Installation in Cube requires additional hardware
Hits: Good at processor
intensive tasks, speedy processor caches. Very good stability. Least expensive upgrade in its class. Can be tweaked to get best performance. 3 year warranty
Misses: Does not work in the Cube.
Requirements: Power Mac G4 (AGP Graphics), Power Mac G4 (Gigabit Ethernet), Power Mac G4 (Digital Audio), Power Mac G4 (QuickSilver 2001), Power Mac G4 (QuickSilver 2002), Macintosh Server G4 (450 & 500 MHz). OS 9.2.1, 9.2.2 or OSX, and PowerMac Firmware version 4.2.8 for all non Quicksilver models
Rating:
(5 possible)
Product: PowerForce Series 100 G4/1.2GHz/1.2GHz/256K/2 MB
Hits: Good at processor intensive tasks, speedy processor caches. Very good stability. Better price/performance ratio than Sonnet.
Misses: Two year warranty not as robust as the other two manufacturers.
Requirements: Power Mac G4 (AGP Graphics), Power Mac G4 (Gigabit Ethernet), Power Mac G4 (Digital Audio), Power Mac G4 (QuickSilver 2001), Power Mac G4 (QuickSilver 2002), Macintosh Server G4 (450 & 500 MHz), Cube. OS 9.2.1, 9.2.2, or Mac OS X.
Rating: (5 possible)
"Real World" Tests
The tests below are from our suite of
real world application tests. These tests feature a
diverse selection of applications commonly used by the
Mac community. The test suite was designed to render
an accurate and well rounded picture of the upgrade's
performance. All of the tests below, were timed with
a stopwatch. The times were then converted to percentages,
relative to our test machine, the Power Mac G4/400, which is
set to 100%. For all scores, higher numbers are better.
Also included, for comparison purposes, is a stock Power Mac single processor G4/1 GHz, which is the currently shipping low-end Tower (January, 03).
The cache setup for all the machine and upgrade options below
are as follows:
Power Mac G4/400 : 1MB of L2 backside cache @ 200 MHz
(half processor speed), (used price: $550)
Sonnet Encore/ST G4/1.2GHz: 256K of on chip L2 cache @ full processor
speed and 2MB L3 cache (price: $530)
GigaDesign G-celerator G4/1.25GHz:
256K of on chip L2 cache @ full processor
speed and 2MB L3 cache (price: $399)
PowerLogix PowerForce Series 100 G4/1.2GHz: 256K of on chip L2 cache @ full processor
speed and 2MB L3 cache (price: $449)
Power Mac G4/1.0 GHz: 256K of on chip L2 cache @ full processor
speed and 1MB L3 cache (price: $1,499)
Desktop Tests
The time between when the circling clock appears and the hard drive shows up on the desktop. Our stock machine (in bold above) took 42.25 seconds and the fastest configuration took 25.31 seconds
Our stock machine (in bold above) took 27 seconds and the fastest configuration took 14.27 seconds
Mac OS X's new File Search function relies on both processor and drive performance. In this test two folders with thousands of files are searched on the terms of 'a, e and o'. Our stock machine (in bold above) took 56.92 seconds and the fastest configuration took 30.43 seconds
Our guess is that the lager L3 cache of the upgrades is causing some fall off in performance ... but we are not sure why this would be. In this test a 60 MB folder with thousands of items is duplicated in the Finder. Interestingly during the test, the stock configured machines used about 30% of the available processing capacity. In the G4/400 Tower, upgraded with a 1.2 GHz card, only 5 - 7% of the processor was being used. Our stock machine (in bold above), carrying out this test, took 20.59 seconds, and the fastest configuration took 10.65 seconds
In the single file test, the the processor & caching capability of the machine is less of a factor. Again, in this test, the stock machine and the Power Mac G4/1 GHz were using 15% of their processor. The upgraded machine 5 - 7%. Our stock machine (in bold above), carrying out this test, took 14.06 seconds, and the fastest configuration took 8.51 seconds
The test above creates and destroys 1,000 windows. See the Let1kWindowsBloom site for more info. Our stock machine (in bold above), carrying out this test, took 67.74 seconds, and the fastest configuration took 32.93 seconds
There is obviously some sort of bottleneck keeping the faster configured machines from performing better. In all configurations 100% of the processor was being utilized. We have noted that on Dual processor machines, running at similar clock speeds, only 50% of the processing capacity of those machines was utilized, during this test. Our Stock machine (in bold above), carrying out this test, took 25.38 seconds, and the fastest configuration took 23.81 seconds
The time Photoshop takes to open is gauged. Our Stock machine (in bold above), took 28.74 seconds carrying out this test. The fastest configuration took 15.27 seconds
Large Document & Database Type Tests
A Macro (series of complex actions) was run in Microsoft's Word on a 500 page document . Word is part of Office X. Our Stock machine (in bold above), took 178.88 seconds carrying out this test. The fastest configuration took 62 seconds
A Macro (series of complex actions) was run on a 10 MB spreadsheet using Microsoft's Excel spreadsheet program. Excel is part of Office X. Our Stock machine (in bold above), took 94.81 seconds carrying out this test. The fastest configuration took 47.21 seconds
This test takes place in a large AppleWorks document. The lack of a L3 cache in the G4/400 Tower, hurts the performance of our test machine, (in its stock configuration). Our Stock machine (in bold above), took 67.07 seconds carrying out this test. The fastest configuration took 17.24 seconds
Number Crunching & Rendering Tests
The Fractal program has been highly tuned to take advantage of the G4 and is precisely the type of work that the G4 was made for. It will also gobble up whatever processing capability is present. This is a good test for assessing the fundamental processing potential of each configuration. Really highlights the processing advantage the G4 processor, when applications are tuned to take advantage of its features. You can download the Fractal program and run it on your own machine to compare scores. You will need to set the fractal to maximum count of 65536, to be consistent with our testing. Our Stock machine (in bold above), took 141.71 seconds carrying out this test. The fastest configuration took 41.95 seconds
In this test, an Aged Effect is run on the 175 MB DV file with the Effect left at its default setting This is a processor intensive task. Our Stock machine (in bold above), took 668.85 seconds carrying out this test. The fastest configuration took 276.84 seconds
A 5MB movie is converted to a QuickTime movie Our Stock machine (in bold above), took 63.09 seconds carrying out this test. The fastest configuration took 27.06 seconds
The 5 MB Movie is converted to a DV Stream ready for import into iMovie Our Stock machine (in bold above), took 482.20 seconds carrying out this test. The fastest configuration took 152.70 seconds
BareFeats Photoshop Test Suite - mimics a production type workflow in Photoshop. If you would like to run this test on your machine, and you don't have a copy of Photoshop, download it from Adobe. Then download the action file and drop it on the Photoshop icon. Our Stock machine (in bold above), took 125.08 seconds carrying out this test. The fastest configuration took 43.68 seconds
Certain of the functions of Photoshop can take special performance advantage of the G4 processor. This suite of filters was run on a 21 MB file. Our Stock machine (in bold above), took 56.96 seconds carrying out this test. The fastest configuration took 22.60 seconds
The test above consists of Photoshop filters and functions that do not take any particular advantage of the G4. Again we used the 21 MB file. Our Stock machine (in bold above), took 239.97 seconds carrying out this test. The fastest configuration took 89.13 seconds
Encoding/Decoding Tests
CD-ROM drive speed and CPU processing power are the factors influencing the scores above. Our assumption is that the CD drive speed is the bottleneck in the upgraded machine. We also think, here again, that the large L3 cache on the upgrades, is helping to suppress scores slightly. Our Stock machine (in bold above), took 240 seconds carrying out this test. The fastest configuration took 75.61 seconds
This is straight processor intensive work and does not take any advantage of the G4 processor... faster clock speed means faster performance. Our Stock machine (in bold above), took 19.72 seconds carrying out this test. The fastest configuration took 8.27 seconds
Multitasking
MP3 Encode, AppleWorks search & replace and folder copy are all carried out at the same time. Scores are flattened out because of the poor MP3 encode performance of upgraded machine. Our Stock machine (in bold above), took 184.05 seconds carrying out this test. The fastest configuration took 69.44 seconds
QuickTime DV conversion and iMovie Effects render are both carried out at the same time. This is heavy-duty processing work. Our Stock machine (in bold above), took 1671.71 seconds carrying out this test. The fastest configuration took 606.42 seconds
QuickTime Encode and Fractal render are both carried out at the same time. Our Stock machine (in bold above), took 207.42 seconds carrying out this test. The fastest configuration took 69.58 seconds
The MP3 Encode, Photoshop Normal and Word Macro tests are all run at the same time. Our Stock machine (in bold above), took 779.44 seconds carrying out this test. The fastest configuration took 167.95 seconds. Again, the faster CD speed of the Power Mac G4/1 GHz pushes it way ahead.
Gaming
Giants is tuned to take advantage of the processor. It also helps to have a better graphics card. Our Stock machine (in bold above), turned in 11.3 frames per second, carrying out this test. The fastest configuration turned in 25.2 fps
4x4 EVO2 is a demanding racing game. Our Stock machine (in bold above), turned in 8.5 frames per second, carrying out this test. The fastest configuration turned in 26 fps
Copyright 1996-2007 by Cider Press Publishing LLC all rights reserved. MacReviewZone is not authorized, sponsored, or otherwise approved by Apple Computer. Apple, the Apple logo, Macintosh, iPod, iBook, iMac, eMac, and PowerBook are registered trademarks of Apple Computer, Inc. Additional company and product names may be trademarks or registered trademarks and are hereby acknowledged.
For
Great Prices On Upgrades Check The Quality Vendors Below
(5 possible) 
