has anyone had success with throttlestop dual ida on an x61t (w/ middleton bios) L7500 CPU

for some reason I adore the 1400 x 1050 resolution on a 12 inch laptop. It's actually quite easy to read unless I'm on the computer for over 16 hours. I have it maxed with an ssd, middleton bios, 8gb ram, intel 6235 with bluetooth 4.0 but I could really use the extra 200mhz on this thing. I just can't get the 9x multiplier with the middleton bios and throttlestop.

 

I know this is not an answer to your question, but:

Having owned X6* tablets with L7400, L7500 and L7700 CPUs - the last one being the fastest available option by the same 200 MHz that you'd like to gain - I've never noticed any difference in speed amongst them.

My route - given that you've maxed out the hardware - would be to start tweaking the software to gain some movement that way....

My $0.02 only...

 

Hello and sorry for the lag...

I've had success under Linux, but the x61t appears to be a bit weird. Perhaps the author of ThrottleStop will know more about this, since it's all the same registers...

My x61t is loathe to go into dual-IDA mode but it will do it if I'm persistent. This is on a regular Midleton BIOS and an L7700, so my max non-IDA FID/VID is 0x919 and my IDA FID/VID is 0xa24.

What I see here is:

1) The IA32_PERF_STATUS only occasionally lists the 0xa24 IDA FID?VID pair as the 'current maximum' on one core or the other (never both that I've seen)

2) will only list the FID when IDA is already enabled but not being used, and

3) apparently will only enter dual-IDA when you happen to toggle the IDA bit when the desired FID is advertised on one of the cores.

I don't know if this is due to to the behavior of the C2D or the BIOS interfering somehow. (I do know I have to turn off SpeedStep in the BIOS, else the Linux kernel will also interfere with the registers, regardless of what I set scaling to. With SpeedStep disabled, Linux will not activate the relevant MSR support)

I set up a little script here that requests 0xa24 then just toggles IDA off then on via 0x1A0 once a second until the request sticks; it can take anywhere from 2 to 20+ tries. Once dual IDA is set, the processor stays at the top IDA multiplier on both cores until you explicitly turn IDA off, at least according to IA32_PERF_STATUS.

BTW, stressed at 0xa24, the CPU sits just under critical temp at full fan speed (about 95C), however it also appears to be stable when substantially undervolted. I don't think I'd run things stuck in IDA mode unless I was undervolting as well. There's just too little thermal margin if ambient was any warmer; no sense setting an OC if it just triggers thermal throttling.