Tyan S1598 Trinity MVP3 ATX

The Bad

Looking around at the competition, Tyan honestly didn't have to do much to make the S1598 a superior board to most others available on the market today.  While that isn't necessarily a bad thing, in this case, it is.  The Super7 motherboard industry, as a whole, has been extremely disappointing.  While some may argue that the K6-2 is by no means a good CPU solution, the K6-III undoubtedly is.  For a user that isn't a hard core gamer, the K6-III delivers very strong performance across the board in integer applications.   Even under games, with 3DNow! support growing in both games and drivers, the K6-III isn't as bad of a gaming solution as the K6-2 was towards the start of its introduction.  The problem?  Most Super7 motherboards out there are very poorly made. 

The S1598 improves on the overall quality we've come to expect from Super7 motherboards, however the true potential of the board is left untapped.  The 686A South Bridge could easily support an AMR slot, and getting rid of at least one of the two ISA slots on the board in favor of adding a sixth PCI slot would be a definite plus for users interested in future expansion.  

The 4ns L2 cache on the S1598 is very impressive, but what isn't impressive is the 8ns L2 Tag RAM that would definitely hinder any overclocking the 4ns L2 cache would make possible, and if that weren't the worst part, as far as overclocking past the 100MHz FSB is concerned, the S1598 doesn't do it.  Especially for K6-2 users that are limited by the 100MHz frequency of their L2 cache, a 112MHz FSB setting at the least would be a very welcome option, however Tyan's young roots in overclocking kept those options out of the feature list for the S1598.

The biggest complaint we had with the S1598 was the VT82C598AT North Bridge used on the board.  Hopefully the sample AnandTech received was the only one to feature this North Bridge as the 598AT was discontinued and then remanufactured under the VT82C598MVP with a few major bug fixes and pin changes.  If you remember back to the days of the i740 incompatibility problems with the MVP3 chipsets (revision CD of the 598AT to be specific, revision CE corrected the problems) you'll remember that the chipset revision to stay away from was the CD revision of the VT82C598AT, the same exact chip used on the S1598 we received.

Although the board worked fine with the Voodoo3 we tested it with, a few hiccups and problems reminiscent of chipset incompatibilities seemed to arise as soon as we tossed in the TNT2 which takes full advantage of the AGP 2X specification.  We're going to give Tyan the benefit of the doubt and assume that the 598AT was just a honest mistake, and that the 598MVP chipset was supposed to be present on the S1598 but to be on the safe side, you should definitely check with your vendor to make sure you're getting the most up to date board.  There is absolutely no reason you should have to settle for a chipset that has since been re-released after a few bugs were discovered.  Shame on Tyan.

The S1598 definitely had the potential to grow into a killer motherboard for the Super7 market, but it seems like shortcomings are to be expected from motherboard manufacturers when it comes to making Super7 boards.  Hopefully manufacturers will take the Athlon a bit more seriously than they did the K6-x series of processors, only time will tell.

USB Compatibility

• Number of Front Universal Serial Bus Root Ports: 0

• Number of Rear Universal Serial Bus Root Ports: 2

• USB IRQ Enable/Disable in BIOS: Yes

• USB Keyboard Support in BIOS: Yes

Recommended SDRAM

Recommended SDRAM: Mushkin SEC -GH PC100 SDRAM; Memory Man SEC -GH PC100 SDRAM
SDRAM Tested: 1 x 128MB Mushkin PC100 SDRAM; 1 x 128MB Memory-Man PC100 SDRAM; 1 x 256MB Corsair PC100 SDRAM DIMM (for compatibility testing only)

Manufacturer: The Memory Man
Purchase Web-Site: http://www.memory-man.com

Manufacturer: Mushkin
Purchase Web-Site: http://www.mushkin.com