Video Board Victory

I recently acquired a copy of Norton SystemWorks v1.0. This provided the opportunity to test the household's Macs for speed, and very interesting it was!

SystemWorks provides a "Systems Ratings" utility to test the host Mac, together with reference system results from a number of standard configurations by way of comparison. This utility provides test figures for five parameters:

• Overall system speed
• CPU
• Video
• Disk
• FPU

Norton's standard testing environment is as follows: Disk Cache at 128K, AppleTalk off, File Sharing off, Virtual Memory off, and Video output at 256 levels.

Rather than set each of the household's Macs to Norton's "standard" configurations, I tested them in the configuration in which they are normally run and compared the results to one sample Norton configuration for a "sanity check."

The household Macs are as follows:

The 9500 had just been upgraded from its original 120 MHz configuration to the shattering power of the G3/350 - well, it's all relative! The original hard drive was upgraded some time ago, and a second video card had been installed. It powers one Sony 15-inch monitor which is attached to the original ATI card and a 22-inch Mitsubishi Diamond Plus 200 monitor is driven by the ixMicro board. Printing is via a LaserWriter 16/600, and it thus requires a network connection, which is currently made by PhoneNet cabling.

The 7500 is running in a 132 MHz configuration with upgraded processor, a new hard drive, and an ixMicro video board. It has a 17-inch Sony monitor connected to the ixMicro board, but the existing 4 MB of VRAM was left in place. As it prints to a Personal LaserWriter LS, file sharing and AppleTalk are both off.

Currently these machines are not networked together, but an ethernet switch has just arrived, so that's the next project!

The 7200 was acquired as a non-working 7500 and retains the original hard drive. Given that the problem was a failed motherboard and I wanted a cheap unit for Internet browsing only, it seemed reasonable to assume that as a 7200 could be upgraded to a 7500, the reverse process could also work. It has. The 7200 mother board came from Shreve Systems for $29 new. Unfortunately, installing Level 2 cache in this unit makes it wildly unstable for reasons which I have never been able to determine, so it runs without one. And, yes, I've tried the CUDA switch and all other known remedies. It supports an old Apple 14-inch Colour Display from the ixMicro board. There's no printing for this one.

Apart from the 7200 motherboard, all the other upgrades have been sourced from Other World Computing, which I find to be a very satisfactory supplier. (I have no connection with this company except as a customer.)

On to the results. I tested each machine as shown above, but with two tests for each (with and without the additional video card) so as to establish the real value of the ixMicro cards. The values are all compared to a Power Mac 6100/60, which scores 100 in every test under Norton's standard configuration.

Compared to Norton's reference scores for the 7600, the 7500 runs a little faster - as would be expected. The 9500 really flies, and the 7200 crawls appropriately, so we seem to have valid testing results. The display scores will obviously be lower than expected compared to Norton's scores because of the larger number of colours being displayed. Also, the 9500 will be hampered by having AppleTalk and File Sharing enabled, compared to the 7500 and 7200, but that's how they are used in my "Real World."

Let's take the 9500 first. The CPU runs nearly three times faster than the 7500/132 - look at the difference in CPU and FPU scores - and it feels like it. The hard drive is around 25% faster, too. But the real interest is the difference in video card performance. The ixMicro card runs at 265% the speed of the original ATI card, and this results in an overall system speed increase of over 50%! Not a bad performance boost for less than $40.

The 7500 shows an improvement of around 225% when comparing the ixMicro card to the internal video, translating to a gain of 30% in overall system performance. What this machine needs now is a processor upgrade.

And the 7200? Well, it just crawls. The stock hard drive is glacial, the built-in video worse at 1 MB (although it would be faster with 2 MB installed). Interestingly, the video speed boost from the ixMicro card is only about 60% in this case, translating to a mere 12% improvement in overall system performance. Shame that it won't support Level 2 cache, as that would certainly speed up matters, but it's running much faster than it did with a dead motherboard.

Compared to my expectations before testing, the differences in CPU are not surprising, while the differences in hard disk performance are greater than I had expected. Of course, your mileage may vary.

I had installed the ixMicro cards in the belief that they would provide an overall speed increase at low cost, but the performance increase in the 9500 and 7500 is way beyond my expectations! On a "bang for the buck" basis, the video card upgrade is clearly a more cost-effective way to improve performance than a CPU upgrade.

Now that was a surprise.