Motherboard Specifications

Back to top

The Good

Shuttle has been recognized numerous times for producing products that excel in quality, performance, as well as the features they provide the user with.  Following in the footsteps of its ancestors, the newly released Shuttle HOT-603 (more or less) boasts excellent quality, performance, and definitely a knockout feature set.

The most highly anticipated Socket-7 Chipset of this year was most likely the AMD640 chipset, unfortunately manufacturing problems and conflicts between AMD and the original equipment manufacturer of the chipset VIA caused the actual implementation of the chipset to be delayed for quite some time.  If you aren't already aware of the fact, the AMD640 chipset is nothing more than an OEM version of the highly acclaimed VIA Apollo VP2 chipset, which is at the heart of the Shuttle HOT-603.  The AMD640 chipset, seen here in all of its glory, provides for the HOT-603's "Pentium II like" feature-set.  This amazing feature-set includes, a cacheable memory area of 512MB (equivalent to the cacheable memory area on a single processor Pentium II system), support for SDRAM, support for UltraDMA Hard Drives with a theoretical transfer rate of 33.3MB/s, support for Linear Burst mode for use with Cyrix/IBM processors, and advanced power management support identical to that found in Intel's TX chipset.  And unlike some other "competitors of the Intel chipset" the performance of the AMD640 (or the VIA VP2) is more or less on par with that of a TX based motherboard.  The 1MB of onboard L2 cache gives the Shuttle HOT-603 the extra edge over the competition, especially when using > 96MB of RAM in memory intensive applications such as Photoshop or NT environments.  Although the AMD640 theoretically supports up to 2MB of L2 cache, Shuttle doesn't provide for this with the HOT-603 making 1MB the absolute limit...ah well =)

The layout and manufacturing quality of the HOT-603 isn't the best it could be, however it is above the "average" for a product of this caliber.  Shuttle is the type of manufacturer that won't stop at the bare minimum when producing a motherboard, this philosophy is evident with the HOT-603.   For an AT motherboard Shuttle managed to pack the 603 with 4 PCI slots and 3 ISA slots in addition to the 4 SIMM and 2 DIMM slots.  Yes, that's TWO DIMM slots!   Although the manual and Shuttle's website mentions that the HOT-603 comes with 3 DIMM slots the model that was tested was only equipped with 2 DIMM slots, which could be a downside if you're looking to use above 128MB of SDRAM with the 603.  The classic "ABIT-style" triple switching voltage regulator setup found its way onto the HOT-603 which is very peculiar considering most TX Shuttle motherboards are missing this feature.  All three heatsinks easily dissipate from the physical regulator which adds to the 603's incredible stability when normally clocked and even when overclocked (to some extent).

Shuttle's decision to include an ATX power connector in addition to an AT power connector gives the HOT-603 a slight layout advantage over its FIC counterpart and also makes it more versatile than the PA-2007 via the uncapping of one jumper to enable support for ATX power supplies. Early reports on the HOT-603 indicated problems with certain power supply units however in the tests conducted I didn't experience any problems with the HOT-603 which could be caused directly by an incompatibility with the power supply of the case (the test case was an InWin G500).   The easy to configure jumpers on the HOT-603 make switching between bus speeds and clock multipliers almost instantaneous.  Shuttle includes and uses a few of their trademark 'jumpers with handles' in the HOT-603 package which make accessing and changing the Voltage and Bus Speed settings a breeze.  The clock multiplier is controlled via three easy to access jumpers located close enough to the CPU to be recognized, but far away enough to make capping them simple. 

Like most newer Shuttle motherboards, the HOT-603 doesn't come with a bulky manual one must thumb through in order to find a few jumper settings, instead it comes with a multi-part documentation system.  Included with the HOT-603 package is a quick reference card, found in other manufacturer's products such as MTech, which completely documents the proper CPU settings (including the 75/83MHz bus speed settings) as well as other useful settings (including the 2.1v core voltage setting for use with the upcoming AMD K6/266 and 300 which the HOT-603 fully supports courtesy of its > 4.0x clkmul ability).  The second part of the HOT-603's documentation is the bundled CDROM which contains such items like the full manual in PDF format, a Shuttle screen saver, as well as useful patches and BIOS updates not only for the HOT-603 but for every single board in Shuttle's recent line of products. 

The setup and configuration of the 603, being a non-Intel chipset based motherboard, is easier than one would expect.  In fact, the configuration of VIA's Bus Mastering EIDE drivers is easier than the horribly incompatible and erratic Intel BMIDE 3.01 drivers.  For those of you having problems finding the drivers, you can download them here.  Upon the initial bootup, the AWARD PnP BIOS Setup utility can be invoked and comfortably configured to the maximum memory timings without many problems if any, the Cache timings can even be taken down to 2-1-1-1 which was nearly impossible with the FIC PA-2007.  The performance of the HOT-603 is excellent overall, making it the new holder of the Best Overall Socket-7 Motherboard Award on Anand's Hardware Tech Page.  With the AMD K6, the performance is unmatched by most competitors and is among the best of the best scoring well above the average in most cases.  The weak-point of the HOT-603 is definitely its Pentium MMX performance, however even with a Pentium MMX 233 clocked at 262.5 the 603 manages to pump out a Business Winstone 97 score of 57.0.  Due to the AMD640's support for Linear Burst Mode, Cyrix/IBM 6x86MX users (as well as older 6x86 users) will be able to enjoy the comfort of an extra 2 - 3% increase in performance as a result of support for this mode.  

For the most part the HOT-603 is an excellent motherboard, however it does have its own share of flaws to deal with.

Aside from the fact that the HOT-603 is only available in an AT form factor the HOT-603 is plagued by yet another problem, unfortunately much more severe than the problems associated with the standard AT form factor.  The HOT-603, unlike the FIC PA-2007 does physically support the 83.3MHz bus frequency, however its stability at that setting is another question.  Before  actually testing the HOT-603, many sources had informed me that there would be problems with the 603 and the 83.3MHz bus speed setting.  The problem as a whole isn't related to one specific component of the motherboard or the chipset, but more or less a combination of minor flaws which together make up for the erratic instability at only the 83.3MHz bus speed (the 603 works almost flawlessly at the 75MHz bus speed).  Obviously, the most viable candidate for a processor that would work at the 83.3MHz bus speed would be the Intel Pentium MMX.  Not only is its voltage requirements much less than that of AMD's K6 and Cyrix's 6x86MX, but it has been known to be possibly the best Socket-7 chip for overclockers since often times the 233MHz unit can be taken up to a blazing 290.5MHz.   And with the Shuttle HOT-603, the Pentium MMX does in fact break the 75MHz bus barrier with the 233 in the test bed able to be run at the 208 setting (83.3 x 2.5) fairly reliably.  Why is it that the Pentium MMX can complete a full Winstone run at the 83.3MHz bus speed when the AMD K6 and Cyrix 6x86MX can't?  Quite a few factors come into play when diagnosing this problem.

To the left you can clearly see the plentiful capacitors surrounding the Socket-7 IC of the Shuttle HOT-603 motherboard, however the 83.3MHz bus speed problem doesn't lie with the capacitors...rather, next to the capacitors.  If you look at the very top of the picture you can barely make out the shape of two of the three "ABIT-style" switching voltage regulator heatsinks, that is where the problems begin.  Remember the hype about how the HOT-603 wouldn't work on some cheaper power supplies?  The HOT-603, could possibly suffer from an inferior voltage regulator setup which would explain why it: has problems with some cheaper power supplies has problems when overclocking processors which draw great amounts of current

The best example of this would be the AMD K6 since it draws an amazing amount of current and has been known to put a LOT of stress on cheaply manufactured voltage regulators.  The voltage regulators on the HOT-603 are in no way "cheaply manufactured" but then again they don't measure up to the quality found in other motherboards like ABIT's TX5N or AOpen AX5T-3.  The K6 worked perfectly at all speeds under 233MHz (excluding settings using the 83.3MHz bus speed and including the default 233MHz setting) however at 250, and 262.5 (not to mention 290.5 which is virtually impossible with the K6) the system could not complete a Winstone test run.  Normally I would attribute this to heat, however the test motherboard was running out of its case at an abnormally cool room temperature in a well ventilated environment.  The CPU was cooled with a 1.5" high heatsink with a heavy duty ball-bearing fan and a small amount of thermal paste was applied between the CPU and the heatsink.  The only explanation for this anomaly would be that the power drawn by the K6 is simply too great for the voltage regulators on the board to cope with, since in most cases where the K6 would fail Winstone the system would spontaneously reboot instead of simply generating an application exception error.  Now although that may explain why the AMD K6 can't be overclocked too easily to clock speeds greater than 233MHz it doesn't explain why the K6 would fail to complete the Winstone 97 tests at 208MHZ (83.3 x 2.5).  The key difference in these two situations is that when Winstone failed at 208MHz it generated an application exception error and did NOT spontaneously reboot.  The problem here directly correlates to the instability with the 83.3MHz bus speed others have been reporting and not the voltage regulators on the HOT-603.  At first I assumed that it would be an SDRAM problem, however after trying 3 different types of SDRAM as well as a pair of Micron 50ns EDO SIMMs I came to the conclusion that regardless of how I paired the memory the 83.3MHz bus speed problems would not go away although I should add that SDRAM based on NEC chips worked seemingly better than those based on other chips.  The next obvious solution to the problem would be memory timings in the BIOS Setup...however even with the most generous settings the instability problems existed in one form or another.   Suggestions were made to add extra cooling to the chipset itself, however I quickly came to the conclusion that the problems didn't stem from the physical cooling of the chipset.  All of this makes you wonder if there was a reason FIC didn't include a 83.3MHz bus speed option with the PA-2007.  It still seems doubtful that the chipset is at fault for the instability at the 83.3MHz bus speed, however rumor has it that Shuttle will be releasing another BIOS update for the 603 which may or may not fix these problems.  The second BIOS revision for the HOT-603 does increase the stability of the 603 with the 75 and 83.3MHz bus speeds noticeably over the original revision, however still not enough to make running on the 83.3MHz setting a realistic option.  I would consider the HOT-603 as not being a contender for the 83.3MHz bus market until Shuttle can fix these problems.  In spite of these problems the HOT-603's excellent performance, quality, and stability (for the most part) earns it the Anand's Hardware Tech Page Award for Best Overall Socket-7 Motherboard.  

IRQ Usage

• Allows user to individually set IRQs for each Legacy ISA card

• Maps PCI IDE IRQ to PCI/ISA

• Auto-detects PnP Cards after HDD Detection

BIOS Settings

Here are my Recommended BIOS Settings for those of you who have been having problems with the HOT-603 and higher bus speeds. Below are my recommended settings for bus speeds < 66MHz, and 75MHz. If you are using Non-EDO RAM (FPM DRAM), then use the settings the in the 2nd (Non - EDO) column, if you are using SDRAM/EDO RAM with a 50/60/66MHz bus speed use the 3rd (66MHz Setting) column. The next column should be used if you are using the 75MHz bus speed and EDO/SDRAM. Finally, if you are looking for the safest and most stable setting, use the last column. There are more Chipset Features Setup options than those I have listed here, but they pertain to Serial/Parallel Ports and have nothing to do with memory timings so I left them out intentionally.

Recommended SDRAM

This little addition to my review layout was put in here just so you all can have an idea of what brand of SDRAM I recommend and have tested with the board, just to avoid problems in the future if you decide to purchase the board.

Recommended SDRAM: Advanced Megatrends SDRAM
SDRAM Tested: 2 x 32MB Advanced Megatrends SDRAM DIMMs

Manufacturer: Advanced Megatrends
Purchase Web-Site: http://www.megacom.com

In recent times, choosing a motherboard cannot be completely determined by a Winstone score. Now, many boards come within one Winstone point of each other and therefore the need to benchmark boards against each other falls. Therefore you shouldn't base your decision entirely on the benchmarks you see here, but also on the technical features and advantages of this particular board, seeing as that will probably make the greatest difference in your overall experience.

Test Configuration

Windows 95 Performance of the Shuttle HOT-603

The Final Decision

Overall the Shuttle HOT-603 is an excellent motherboard, and if you're not too interested in the overclocking possibilities the 83.3MHz bus speed gives you then it would seem to be the perfect motherboard...for now that is.  Later this month some of the biggest names in the industry will be flaunting their newest motherboards based on the VIA VP3 chipset (the successor to the VP2 and AMD640 which support all of the features of the previous two and adds support for 1GB of cacheable memory as well as AGP) which could very well replace the HOT-603 and the PA-2007 at the top of the list, but for now if you must have a motherboard give the HOT-603 a good look over since it does have quite a lot to offer.