Announcing Windows 10 Insider Preview Skip Ahead Build 17639 - April 4 Insider

creative xifi sound card didn't install on this update

I would not download an ISO from those Russian sites ever, for that there exists the mini-server, with which I can download the UUP files directly from the Microsoft servers and create the ISO myself

MUser said:

I agree is all very odd. I downloaded a fresh copy of Inspectre for the test you saw two days ago. I will try poweshell over the weekend. I think its all genuine and I suspect the microcode for the i7-920 has slipped through to build 17639. I see your EE processor has a different CPUID so this may be the reason?

I’m going to make a copy of the file holding the microcode from build 17639 just in case its reverted/changed for the production release of SCU.

All very vexing.

Here is the output from the power shell

johngalt said:

But the other side of it is that the pre-mitigation microcode that was originally shipped did not actually fix the problem, hence the roll back on it. Furthermore, the Intel compliance PDF shows that my CPUID and yours both contain the Core i7 920s - so which is which?

Either way, I don't see why the 965 EEs have a different CPU ID from the 975 and 980, which are grouped with the majority of the rest of the original generation Core i7s. So, I really don't understand what is going on here.

I know it is not the architecture, because the motherboard I have , which has the LGA1366 socket, can also accommodate that 975 and 980 - in fact, IIRC, that entire family of Bloomfield CPUs were all LGA1366 CPUs, and set up for triple channel memory (again, if memory serves me). So, I don't see why the 965 was partitioned off with a different CPUID, but then when the 975 (and then 980) were developed, that Intel would dump them back onto the older CPUID.

That, in fact, is what made me begin to doubt the veracity of the April compliance PDF in the first place.

Here is a comparison of the three: Intel Product Specification Comparison

I am not so sure about "rolled back", they said "stopped" due to one or more of three reasons. It could just be lack of commercial take-up (by motherboard makers to be bother updating BIOS) or no big single customer or group exposed (e.g government or cloud supplier). It does not mean the microcode was/is faulty? I am presuming it is good microcode for the i7-920, and its test out ok for me, but Intel decided to lump it with others based on the high level generational name and to not officially release it but this change came too late for Windows 10 build 17639? But as you highlight we have no clarity!

There is an i7-920XM too, which I do not have, and I wonder if this is the root of the confusion in the Intel documents. Typos may be?

I know my i7-920 was a very early build, I pre-ordered it from Dell and it shipped on 15/8/2009. Here is some further details from Dell/PC-Doctor (I presume the "h" suffix indicates hex in this case and not a variation?):

Name Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz, Cores 4, Threads 8
Signature 106A5h, Family 6h, Model 1Ah, Revision D0, Technology 45 nm
Max Qualified Speed 2.67 GHz, Max Turbo Boost Speed 2.93 GHz, Bus Speed 133.00 MHz
Features MMX, PSE36, EM64T, SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, XD, VMX, IDA

I have run a few tests on it (i7-920) with build 17639 and all is fine, including all the OCCT stress tests, userbenchmark, GeekBench, full Dell/PC-Doctor tests and none show any issues, generally its a smooth running machine.

It will be interesting to see what is released, or removed, when the production release of Window 10 SCU arrived on Tuesday?

MUser said:

Here is the output from the power shell

See microcode update at bottom in the below screen capture (This is a 9 years old i7-920 with no BIOS update since 21/9/2009!!!!).