Home

|
	Please Note: Because the following information is often based on personal speculation, rumors, and unsubstantiated comments from others, please do not consider it to be the absolute truth.

October

10/15/97 - The Cyrix/IBM 6x86L processors are on their way out for good. Authorized distributors are having a very difficult time getting any of these CPUs. Cyrix/IBM stopped producing wafers in January and the supply has finally dried up. If you want one, you better get it now.

August

8/21/97 - Cyrix hit the road this summer. They have been putting on summer reseller seminars around the country. I received some information from a friend who recently attended one. Here are a few tidbits from the seminar for you.

The new 6x86MX PR200 processors have hit the street and are marked 150MHz (2 x 75MHz). At the show a technician said that they have some PR200 166MHz (2.5 x 66MHz) parts, but they are special order and will probably cost more. The upcoming PR266 will initially be a 208MHz (2.5 x 83MHz) part until they get the yields up and then they will move to 225MHz (3 x 75MHz). Cyrix will have a PR300 by year end or early 1998. They haven't decided on a bus speed yet.

The PR233 and up will continue to be 2.9V. Cyrix is looking at producing some 6x86MX parts to run at 2.1V for notebooks.

National Semiconductor fabs will be used to produce the MediaGX chips, while the IBM agreement will stay in place for the 6x86MX. By September the 6x86 will be an endangered species. In fact, they stopped producing 6x86 wafers in January.

The hint is that ALi has a chipset that officially supports 83MHz (ALi is ACER). This is who Cyrix is looking at right now to support their PR266, but VIA may have their 100MHz chipset available by then.

July

7/22/97 - I have been informed that Cyrix/IBM are going to make changes to their Performance Rating system for the 6x86MX because of inadequate 6x86MX manufacturing yields and frequencies.

Here is how the new temporary PR rating is suppose to look:

Cyrix website no longer talks about clock rate or multipliers. Obviously, they are trying to slide the changes in unnoticed! It looks like the rumors are true.

The reason for this change is that Cyrix/IBM can't satisfy the demand for the 6x86MX PR200 chips because they are not able to produce enough parts that can run at 166MHz and above. Cyrix/IBM can better meet the demand if they sell a 150MHz instead of a 166MHz part as a PR200. Cyrix/IBM can and benchmarks prove that they can get PR200 performance from a 6x86MX by clock doubling the 75MHz bus. Personally, I wouldn't even think about seeing a 6x86MX PR233 on the street until late Q3. Please note that this will be a temporary change in the PR rating and Cyrix/IBM will move back to the original model as time progresses and yields improve.

Cyrix/IBM aren't going to hide this new fact, but they aren't going to broadcast it either. This will make the PR rating more confusing than ever (if that is possible), but it has saved their rear-end in the short term. Like the PR233, there will eventually be two different versions of the 6x86MX PR166 and 6x86MX PR200. For example, there will be a 6x86MX PR200 (66 x 2.5) and 6x86MX PR200 (75 x 2).

7/02/97 - Cyrix/IBM do have a small amount of 6x86 parts that yield 83MHz. Someone came to them and said that their chipset was designed to run at 83Mhz and that they were building a motherboard to support it. As a result, don't be surprised if in a couple months you see a fairly well known company with a killer 83/166 6x86 budget system.

June

6/21/97 - Cyrix/IBM are investigating using 6x86 (not 6x86MX) in mobile solutions. Cyrix/IBM can sort the 6x86 to meet Intel mobile voltage specs, but there has to be enough interest from mobile motherboard providers in order for them to do this.

6/19/97 - The next generation of Cyrix processor is codenamed the M3 and will be socket-7 compatible. Intel has stated that the TX chipset will be the last Intel chipset to feature socket-7 support. This is an effort by Intel to migrate the masses to the Pentium II and Slot I architecture where they will be offering the latest and greatest chipset technologies. Cyrix has no future plans to get into the chipset business. Lets all hope that AMD, VIA, SIS, and others pick up Intel's slack and keep the socket-7 a viable option for years to come.

March

3/22/97 - The threat of the AMD K6 looms large and it is going to be released approximately 2 months before the M2. Cyrix and IBM run the risk of losing their hard-earned reputation as the superior alternative CPU to the Pentium in terms of performance in the interim. As a result, the M2 partners are looking at any possible way to accelerate production availability. Don't be surprised if it is released before June.

3/03/97 - Cyrix will never release a software patch to remedy the pre-revision 2.7 6x86 and Windows NT 4.0 performance issue. Did anyone really expect that they would?

February

2/18/97 - Cyrix and IBM plan to release a 6x86 166MHz (clocked-doubled 83MHz) before the end of the year.

January

1/27/97 - This one is a jewel. I never cease to be amazed how these morsels magically end up in my mailbox. When the clock strikes 5:00 P.M and people start calling it a day, I’ve heard that the guys at Cyrix like to wind down by playing some hard core network Quake. I find that pretty ironic considering the Quake players bash Cyrix like no other. Maybe some of the engineers play every now and then and realize that they need to beef up the FPU on the M2.

1/20/97 - The revision 3.7 and above P166+ and P200+ cpus are made from same advanced silicon process. The revision 3.7 and above P166+ overclocks to P200+ quite well. Here is why. Cyrix tests the cpus in a wide temperature range. They have been having problems with the P200+ failing in extreme temperature ranges and thus mark them as a P166+ since its passes the temperature tests at that speed. Since most of us do not run our systems in extreme temperature environments, the revision 3.7 and above P166+ series seem to have a high overclocking success rate.

1/17/97 - The M2 core voltage requirement may be changed from 2.5V to 2.8V to match the 2.8V Pentium MMX. This will allow an easy upgrade for anyone who wants to upgrade a Pentium MMX to a Cyrix M2.

1/10/97 - I received a very interesting email today. An individual that I trust was privy to some correspondence which describes the differences between the Cyrix 6x86 chip revisions that Cyrix will not disclose. He was gracious enough to forward it to me. I decided to put this information on this page rather than the 6x86 FAQ because he asked that I keep the anonymity of the real source. As a result, I could not verify the information with the source myself.

2.4 -> 2.5
Cyrix found a problem in the L2 cache control signal which caused certain motherboards to be unstable. They changed the electrical signal to cope with this problem.

2.5 -> 2.6 and also 2.6 -> 2.7
Inorder to yield more P166+ and P200+ cpus, Cyrix improved the flow of electricity between the layers and modified the location of some units.

2.7 -> 3.7
Cyrix changed the manufacturing process.

3.7 -> 3.71
Inorder to yield more P200+ cpus, Cyrix improved the flow of electricity between the layers.

1/9/97 - Although the M2 partnership with Evergreen was once the real thing, with Jerry Rogers (Ex-CEO) gone and a new management team in place, Cyrix has decided to cancel the $199 M2 upgrade deal.

Has anyone else noticed that the Future's Built In demo and the majority of graphics related to the Future's Built In program have quietly been removed from the Cyrix website. The Future's Built In logo, which was prominently displayed on the majority of Cyrix web pages, is also gone. I wonder if this means that Cyrix is starting to back away from their promise that the future's built in.