|
|
Review date: 4/19/98 | |
|
| |
| Introduction | |
|
MacTell sells a line of G3 CPU card upgrades based on the PowerForce design called the PowerJolt. The review sample they supplied was a 266mhz model with a 1MB backside cache running at the full processor speed (1:1 ratio). As many of you already know, there are currently only two different G3 CPU card designs on the market, the Newer Technology cards with limited speed adjustments and the PowerLogix PowerForce design which offers a wide range of speeds. Although based on the PowerForce design the MacTell card does have two differences that separate it from the rest - slightly faster cache (4ns vs. 5ns) and a unique heatsink that contacts the cache chips as well as the CPU.
As detailed on the hardware specifications page, close examination showed that Mactell had made a change from the norm in the heatsink design - it contacts the cache chips as well as the CPU. The 1:1 (and most 1.5:1) cache speed G3 cards all use the metal lid IBM cache and this is an effort to improve cooling of the chips. Since the 2:1 (slower) cache speed cards use standard plastic or ceramic cache chips, they will probably not have this feature. I'll be honest, I'm not sure the cache really gets that warm in these designs. My initial concern is that the heatsink contact may actually tend to transfer heat to the cache rather than away from it, but in extended use at 300/300 I did not encounter any problems. However I ran that same speed with a XLR8 266/177 card that did not have the heatsink contacting the cache chips so I didn't really see a benefit to the design, but it's possible it could contribute something. I also noted the card used 4ns rated cache, faster than the 5ns parts used most other high end G3 cards. In actual use this did not prove to have any real benefit however. MacTell owners who bought their cards awhile back may want to download the latest PowerJolt v2.01 extension, as it is a later version that the one that shipped on the floppy disk included with the card. The only difference I noted was that it loaded earlier in the extensions list which could be a benefit to compatibility. As with the XLR8 266 G3 card based on the same design, I had to remove the PowerTower Pro 1MB L2 cache dimm to allow the card to boot reliably. While on the subject, the PowerJolt extension did not seem to get along with the ByteMark benchmark, generating a Type 1, 2 or -7440 (?) error halfway through the test when running any speed over the stock 266/266. I did not see these errors with other G3 PowerForce cards I've tested and normally use that as an indication that the card is running at too high a speed to be reliable (as the Bytemark tests exercises the CPU very heavily). During my testing of the PowerForce G3 220 card for instance, backing off a few mhz after seeing this error would result in the test passing. Since there were no other errors (and the fact that without the extension active it ran the test fine) I decided to post the 300/300 results. I also found it odd that the stock PTP 180 also had type 2 errors in Bytemark, although the other CPU cards tested in the same system didn't.
I rated the card using my standard categories as shown below:
Test System: The base system used for test was a PowerComputing PowerTower Pro 180, running OS 8.1 with the standard array of Quicktime 2.5 and Quickdraw 3D 1.5.3, PC exchange, Connectix's Speed Doubler 8.1. Disk cache was 512K, virtual memory was off. Installed Ram was 128MB, and other than a video card, all other PCI slots were empty. For this review the ATI XClaim 3D video card was installed in the lowest slot. To remain consistent with my other CPU card reviews, screen resolution was set to 1024x768, thousands colors. The L2 Cache dimm of the PowerTower Pro had to be removed for the card to boot reliably in my PowerTower Pro (just like the XLR8 266 card).
| |
I welcome comments about this review and/or the product.
| |
|
| |
|
You can follow my preferred path through the review by continuing to the next page, or use the links below to jump to a specific page. | |
|
| |
|
Intro | Benchmarks
| Appl. Tests | Software Controls
| Documentation | Specifications | Summary
- or - | |
Copyright © Mike, 1998. Users of this web site must read and are bound by the terms and conditions of use. | |