Intel and AMD are battling for the most cost effective processor solution for the home user.  Both are claiming outstanding gaming performance, excellent business application speed, and both are arguing over who is faster.  Observing from a distance is none other than the popular low-cost processor manufacturer, Cyrix.  After the initial fame the successful 6x86 processor brought Cyrix, the fab-less manufacturer decided it was time for a more powerful successor.   While everyone expected the successor to the 6x86 to be a much more powerful processor geared towards the high end market while making sure that gamers weren't left out, no one was truly surprised when the long awaited successor turned out to be nothing more than a higher clock speed processor with more L1 cache and MMX instructions.  So went the introduction of the 6x86MX...

Much more of a disappointment than its predecessor, the 6x86MX demanded that the market be given an alternative solution to cover up its shortcomings.  Unfortunately for Cyrix, that alternative came from AMD with the K6.  Simply hating being the last to jump on any bandwagon, Cyrix decided to push the limits of their current microprocessor architecture once again.  Exactly one day before Intel announced the release of their first processors with support for the 100MHz Front Side Bus, Cyrix made a press release about their next generation 6x86MX CPU, the M-II, with support for the 100MHz FSB.  A trick like this has been known to be in Cyrix's mystery bag, tracing back to the days of the original 6x86 when Cyrix pushed the FSB limits past the officially supported 66MHz clock to 75MHz with their PR-200+.

The press release Cyrix made about their new M-II processors was a bit disappointing to most Cyrix advocates that expected a revamped Floating Point Unit and higher clock speeds from the next generation 6x86MX.  The M-II became known as nothing more than a faster 6x86MX with a new name that surprisingly enough resembled the P2 in the simplicity of its nomenclature. 

The M-II, like the original 6x86 and 6x86MX isn't identified by its clock speed, rather by its performance rating number.  For example, the M-II 300 doesn't actually run at 300MHz rather provides business application performance generally greater than or equal to a 300MHz 6th generation processor such as the Pentium II or the K6-2.  In actuality the M-II 300 operates at a 233MHz clock frequency however it can be run at anywhere from 200MHz (100 x 2.0) to 233MHz (66 x 3.5) including 208MHz (83 x 2.5) and 225MHz (75 x 3.0) making the M-II a processor whose versatility is encouraged to be taken advantage of by the manufacturer themselves.   Officially the M-II 300 only supports operation at 233/66 however the above settings will work perfectly fine with the processor as long as the other components in your system have no problem with the increased FSB frequency.

Set apart by its unified 4-way set associative 64KB L1 cache (generally the L1 cache is split equally into two parts, data and instruction set cache) the M-II uses the same exact core found in the 6x86MX processor, meaning you get the same processor with a new name and a higher clock speed.   The operating voltage of the M-II, like previously 6x86MX processors and the AMD K6 166/200, is 2.9v which makes the M-II an option for a great number of users with motherboards a year old.  If your motherboard doesn't support the 2.9v core voltage then you may be out of luck, while the processor can work at 2.8v and 3.0/3.1v doing so isn't encouraged by the manufacturer.  The 2.8v setting will only work if you have a very well designed motherboard (from an engineering perspective), if the 2.8v setting on your board fluctuates beyond a reasonable range then you will probably experience quite a bit of instability.  The same goes for the 3.0/3.1v settings, so your best bet for the M-II would be to get it with a board that officially supports it.

The chip itself bears the new M-II logo on the front but is physically no different than even the old 6x86 chips.  Manufactured using a 0.35 micron die the M-II generates a considerable amount of heat, however if you are familiar with the older 6x86 (not the 6x86L) chips then you don't have to worry since the heat production of the M-II is no where near that of the first 6x86 processors.   A standard heatsink and fan will suffice, heatsink compound will definitely help if you plan on pushing the processor to its limits...in other words, overclocking...

The M-II tested by AnandTech had no problems handling FSB speeds up to 100MHz on Super7 motherboards.  Unlike older 6x86MX processors which simply wouldn't work at the 100MHz FSB, the M-II 300 didn't seem to have many apparent problems with the setting.  Using the FIC VA-503+ Revision 1.1B, which unofficially supports the 112 and 124MHz FSB frequencies, the M-II 300 would consistently crash under Windows 9x and in spite of reliable operation for a limited period of time, the M-II test system would not work reliably enough at those two particular settings to be considered an option. 

Even using a higher core voltage setting wouldn't allow the M-II 300 to pass the 233MHz mark reliably.  The test system would boot at 250MHz (83 x 3.0) but stability, as mentioned before, wasn't solid enough to be considered a viable overclocking option (Winstone crashed 3 of the 5 times run).  The next in the M-II line, the M-II 333, should prove to be much more of an overclocker as it is a 0.25 micron chip which will produce less heat and operate at a lower (2.5v) core voltage setting than current M-II processors.  In comparison to the other Socket-7 processors out currently, the M-II is an extremely poor overclocker as the 233MHz part wouldn't even make it past 250MHz.   Increasing the core voltage on the M-II didn't help too much to increase stability, the sweet spot for the M-II 300 seemed to be the 3.0v - 3.1v region, 3.2v just caused more problems than it created, and anything above that is simply out of the question for a stable system.

Unlike the AMD K6-2, the M-II from Cyrix doesn't require a special motherboard to get the most performance out of it.  The processor will receive a 2 - 5% increase in performance on motherboards that supports its unique form of Linear Burst SRAM addressing.  Motherboards based on non-Intel chipsets generally support Linear Burst Mode, such as those based on VIA chipsets, support for this feature isn't required but it is a nice plus for owners of motherboards that support the setting. 

Weaknesses

Strong business application performance is what makes the M-II the chip it is, and not an odd looking 486.  An ever present weakness of Cyrix processors has been their extremely weak FPU performance, meaning 3D gaming performance on an M-II is pretty much unacceptable in comparison to the competition's performance.  Because of this poor FPU, the M-II has gained a poor reputation among gamers and it is the main reason that the reputation of Cyrix processors in general has suffered so greatly. 

Cyrix vs the Rest

How does the M-II stack up to the competition?   Providing business application performance equivalent to a K6 300 or Pentium II 300, Cyrix does offer a lot of bang for the buck.  Considering the M-II 300 can be purchased for just under $100, and considering that it offers performance in general business applications (i.e. Microsoft Word, Surfing the Net, etc...) that is competitive with the rest of the competition, the M-II stacks up quite well.  In essence, you can consider the Cyrix M-II the opposite of an Intel Celeron in that it has outstanding business performance at the tradeoff of poor 3D gaming performance. 

While Cyrix has been boasting that the performance of the M-II is greater than that of the Intel Celeron, you must keep in mind that the Celeron and the M-II are geared towards two completely different markets.   There is no way that the M-II can compete with any of the other 6th generation processors on the market in terms of 3D gaming performance, even when armed with a Voodoo2 accelerator, the performance of the M-II 300 under Quake 2 is barely reaching 30 fps.   The Celeron has an extremely strong FPU and is therefore much better for gaming, whereas the M-II has a considerably weaker FPU but offers business application performance equivalent to that of a Pentium II.

In comparison to a K6, you may want to go ahead and purchase a K6 over an M-II provided that you have the extra cash to make up for the difference in price.  The K6, especially the K6-2, is much more of a well rounded chip when it comes to performance.  Although the raw FPU power of both the K6 and K6-2 is considerably slower than that of a Celeron/Pentium II, the FPU of the K6 is still considerably more powerful than that of the Cyrix M-II making it a better option for users with gaming intentions as well as productive ones for their computer. 

• AMD K6-2 333, Cyrix M-II 300, Intel Pentium MMX 233

• FIC VA-503+ Motherboard

• 64MB PC100 SDRAM

• Western Digital Caviar AC35100 - UltraATA

• Matrox Millennium II AGP Video Card (4MB)

• Diamond Monster 3D-2 Voodoo2 Graphics Accelerator (12MB)

The Pentium II comparison system differed only in terms of the processor and motherboard in which case the following components were used:

• Pentium II clocked at 300MHz (66MHz FSB), 333MHz (66MHz FSB), 400MHz (100MHz FSB)

• ABIT BX6 Pentium II BX Motherboard

The following drivers were common to both test systems:

• MGA Millennium II Drivers v4.07.00.700

• DirectX 6 SDK

All tests were run at 800 x 600 x 16-bit color

Solid business application performance, what more can you expect from a processor whose sole existence relies on strong performance under business applications?

Horrible video playback scores, don't expect the M-II to be great for a DVD system, you would be better off buying an external player for your TV than running a Software Decoder off of an M-II.  The Cinepak, Indeo, and MPEG-1 playback numbers are terrible for a 6th generation processor.   While many will argue that the M-II was only running at 233MHz, you must remember that Cyrix is selling the processor as an M-II 300 and as you can clearly see the performance of the M-II isn't at the level of a 300MHz processor as far as video playback is concerned. 

Are Quake 2 and Turok playable on the M-II 300?  Sure, provided you are willing to spend an additional $220 on a Voodoo2 accelerator to achieve that playability.  This is more proof to the fact that the M-II is not an entertainment processor, games don't mix with Cyrix...not yet at least.   For surfing the net and typing up documents, the M-II is a $95 wonder, for fragging in Quake, it is a wonder that you're considering this $95 processor.

The M-II is in no way a high-end chip, however it is great at what it does, and if all you find yourself doing is typing up documents, surfing the net and playing solitaire then the M-II 300 will last you quite a while.   Even the user that plays the occasional 3D game may want to go with a K6 here, especially since the cost of a lower end K6 has dropped well below the $100 mark.   Once again, it all depends on the type of user you are...but don't complain if your new M-II isn't running Quake 2 as fast as you would like it to.

Cyrix M-II 300 Processor
Vendor: Computer Parts USA
Website: http://www.computerpartsusa.com/