The AMD K5 has been coming slowly, but now it's there and getting better and better. By now, you can get it from a PR 75 up to a PR 166. It's FPU is still slower than the FPU of the Pentium, but it's better than the FPU of the 6x86. It also will soon be released as a split voltage version, which will reduce heat problems of this CPU as well.

Here a table to show the weird CPU speed/bus speed/multiplier relations of the K5 CPUs:

As you can see, the PR120 runs at the same speed as the PR90, the PR100 runs at the same speed as the PR133. These CPUs obviously have got a different die. The PR120 and PR133 are more advanced than the PR90/100. This however only touches the integer performance. The FPU performance of the PR133 is the same as of the PR100.

Due to these strange multiplier settings, the K5 has got a fixed multiplier, which cannot be changed from out side via multiplier jumpers. The PR75 to PR133 will run at 1.5 multiplier regardless what the jumpers say, the PR166 will also always run at x1.75.

When IDT announced the C6 CPU in the first half of 1998 it came as quite a surprise to most of us. After this first introduction it became pretty quiet around IDT again and only now this new Socket 7 CPU starts shipping. Targeted against the Intel Pentium MMX, IDT claimed that the C6 has got about the same integer performance at the same clock speed as the Pentium MMX. IDT was also not making any secret of the fact that the floating point as well as the MMX unit of the C6 will not be as fast as in the Intel Pentium MMX. The C6 is targeted to the lower end market, similar to the lower clocked versions of the AMD K6 and the IBM/Cyrix 6x86MX CPUs. The special trick of IDT is supposed to be the fact that the chip is fairly small as well as requiring only little power and IDT is therefore pretty confident of raising the clock speed to up to 400 MHz soon as well as using the C6 in notebooks where low power consumption is crucial.

The architecture of the C6 is much less sophisticated than the architecture of its competitors AMD K6 and IBM/Cyrix 6x86MX. No 'register renaming' or 'out of order execution' is used here for reaching Pentium MMX performance but simply a large L1 cache in cooperation with a pretty classic but straight forward microprocessor design.

At the Computex in June IDT only showed a few systems with C6 CPUs running at 150 or 180 MHz. I also had one 150 MHz engineering sample a few months ago which couldn't really impress me a lot. Now IDT shipped a whole system to me with a 200 MHz C6 so that I could test if this new Socket 7 CPU is living up to its claims. I compared the C6 200 to the Intel Pentium MMX 200, the AMD K6 166 and 200 and the IBM/Cyrix 6x86MX PR2 166 (running at 133 MHz) and PR2 200 (166 MHz). The price of the C6 200 lies pretty much somewhere in between these CPUs, with the Intel Pentium MMX as the most expensive, the IBM/Cyrix 6x86 PR2 166 as the cheapest.

The IDT WinChip C6 CPU is certainly no particularly great performer, but that was never claimed by IDT either. It is a 'low trouble' CPU with a decent integer performance under Windows 95 that might have some impact at the low end sector of IBM's and Cyrix's 6x86MX market. I personally like this CPU and would prefer it to the 6x86MX PR2 166, because it's so much easier to handle, however as long as the 6x86MX is cheaper, I wouldn't see a sensible reason for spending more money on the C6. Although my expectations were extremely low, the C6 shows that it can beat at least the slowest 6x86MX or offer pretty much the same performance.The C6 will be sold for $135 (200 MHz), $90 (180 MHz). The $135 for the 200 MHz version are pretty much the same what the PR2 200 rated K6 and 6x86MX CPU will cost from beginning of November. Both CPUs are faster than the C6 200, so that IDT will have to lower prices pretty soon if they want to sell the chip.

What I would like to see in the near future is a real low power version of the C6, which is running at 2.8V or even less. Then I hope that IDT will be able to crank up the clock rate yields, which could result in quite a market share, probably taking over customers of IBM and Cyrix. I think that the biggest future of the C6 will be in the notebook market in case they can reduce the voltage to about 2V. This and a clock rate of 400 MHz will even give Intel and its upcoming notebook version of the Pentium II quite a little bit to think about.

The road map doesn't look too bad regarding the above said. IDT is planning a 225 (3x75) MHz and a 240 (4x60) MHz version for this November and next year there will be a WinChip C6+ with improved FPU performance (as fast as Pentium MMX) and much improved MMX performance (faster than Pentium MMX) as well as extended x86 instructions that shall improve 3D performance up to threefold. These new instructions will be distributed within Microsoft's next Direct3D. The WinChip C6+ will use a 1/4 micron die and run at only 2.5 V.

For late '98 we can expect a 300 MHz C6+ as well as a new version with on-chip 256kB L2 cache.

The Cyrix 6x86 or M1, as it was called before, is a strong contender to the Intel Pentium and is able to outclass the Pentium in quite a few respects. This CPU might also be able to push technology forward more than the current Pentium CPUs, because since the 6x86 P200+ was released, a new bus speed of 75 MHz turned up at the horizon to replace the old maximum of 66 MHz of Intel CPUs. This will require new memory technology as well as new chipsets and therefore new motherboards.

This however is also the downside of the new 6x86 P200+, because you can't run this chip in a Triton FX or HX board for they don't support a bus speed of 75 MHz. Only to remind you - the 6x86 P200+ runs at a clock speed of 150 MHz, which is a multiplying of the 75 MHz bus speed with 2. Would you run this CPU with a bus speed of only 50 MHz and a multiplier of 3, you would loose all the speed advantages due to this pathetic slow bus speed.

Technical Specifications of the 6x86

Problems with the 6x86 and the solution

The 6x86 CPU was not a child that came to the world and was making every person, that had to deal with it, happy straight away. The 6x86 had it's problems and Cyrix had to fight with a lot of trouble since the 6x86 release. However this doesn't say that Cyrix wasn't working extremely hard to survive by fixing one problem after the other. The latest improvement the 6x86 has received from its creator is the most pleasing one and eventually now I am happy with the 6x86 as well. The letter 'L' behind the CPU name is making all the difference to me and hopefully to all the possible 6x86 buyers as well.

The 'L' stands for 'low voltage' and doesn't say anything else as that the 6x86 now also has got the 'split voltage' feature known from the newly released Pentium MMX. Now for the Pentium this isn't making much of a difference, since the Pentium never had to fight any serious current or heat problems. For the 6x86 however, this could be the final liberation from his bad name as 'chip fryer'.

The 6x86L has just as the Pentium MMX a 2.8 V core and a 3.3 V I/O interface voltage. Cyrix expects a power reduction of more than 25% of the previous models, which will keep the CPU much cooler and won't ask for more current than a Pentium MMX.

This is probably the solution for all the people who have been scared by the recent problems with the 6x86. Unfortunately it comes a little bit late, since everyone is waiting for the M2 and the Pentium MMX 200 is much faster than a 6x86 P200+ (which is even on a lower level than the Pentium MMX 166 in most applications). However, who thinks that a P200+ is still by far fast enough and who doesn't want to spend as much money as for Intel CPUs, should get a 6x86L and nothing else. Getting an old 6x86 without the 'L' now is considered to be completely crazy, unless the old single voltage 6x86 shouldn't be considerably cheaper than the 6x86L.

I don't think there's too much to say about the Pentium, because there's that much information already being released on the net and in magazines. However there's one important thing, which doesn't seem to be known too well. It's the fact that Pentiums of the same clock speed are not all the same, but can be VERY different, and you should know that before you get one ! There are now three reasons for that to be considered:

The Technical Advantages of the Pentium MMX over the Pentium Classic

The family consists of processors at 150 Mhz and higher and is easily scalable to up to four microprocessors in a multiprocessor system and is suitable specially for servers.

Features

It is simply Pentium Pro + MMX. The Intel Pentium II processors deliver good performance and are fully compatible with existing  software. The Intel Pentium II processor family includes 233MHz, 266MHz, 300MHz, 333MHz, (FSB Speed 66 MHz)  and   350MHz, 400MHz, 450MHz  (FSB Speed 100 MHz) versions. The processor core is packaged in the Single Edge Contact (SEC) cartridge.

	Available in speeds from 233MHz up to 450MHz.
	Includes  MMX technology.
	At 450MHz, delivers a 25% performance boost over the 333MHz Pentium II processor
	Dual Independent Bus (DIB) architecture increases bandwidth and performance over single-bus processors.
	350MHz, 400 MHz  and   450MHz versions improve system bandwidth and performance by increasing the system bus speed from 66MHz to 100MHz. The Intel 440BX AGP set enables the 100MHz system bus to increase peak processor data transfers to the rest of the system by 50%.
	32K (16K/16K) non-blocking, level-one cache provides fast access to heavily used data.
	512K unified, non-blocking, level-two cache.

The Intel Celeron processor, is for  the Basic PC (sub-$1000) market segment. The Intel Celeron processor meets the core needs and affordability requirements common to many new users. Intel Celeron processor benefits from the same P6 microarchitecture core as the Pentium II processor. It is same as P-II minus the l2 cache. Intel was forced to release this as a cost effective soulution because of the competition it has been facing in CPU market. Currently Celeron  processor is available with 128 KB L2 cache.  

Back in end 1998 at the Microprocessor Forum in San Jose, California, the PC-world watched and listened in amazement to Dirk Meyer's first presentation of K7's or now Athlon's architecture. It was quite obvious to experts as well to most other listeners, including Intel employees, that this new AMD processor would mark a new era in the processor world, if AMD could make its promises come true. Now finally, the waiting is over and we can look at a new processor that is indeed living up to all the positive expectations that arose at the end of last year.

Athlon was designed with the future in mind. The architecture is designed for very high clock rates and obviously superior to Intel's P6-architecture, including the Pentium III. If AMD can deliver Athlon and move over to .18µ-process soon, Athlon will be a continuos threat for Intel, even beyond Coppermine. And currently Althon beats Intel flagship chip Pentium-III  in almost all spheres.