Take a look at our
ThinkPads.com HOME PAGE
For those who might want to contribute to the blog, start here: Editors Alley Topic
Then contact Bill with a Private Message

Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

T60/T61 Series
Message
Author
fourthree
Freshman Member
Posts: 98
Joined: Fri Dec 16, 2016 8:15 pm
Location: Bat Yam, Israel

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#271 Post by fourthree » Tue Jul 04, 2017 4:16 pm

Wow, thread is dead..

If the socket needs to be modified, can something like http://www.ebay.com/itm/2-5-10X-The-Thi ... xZ1zZEiCyg work instead?
Dorkstation: T601f
Ultraportable: X61s

el-sahef
Sophomore Member
Posts: 131
Joined: Sat Mar 17, 2007 6:20 am
Location: Frankfurt, Germany

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#272 Post by el-sahef » Wed Jul 05, 2017 2:45 pm

No, not in this case.
T70 15" 4:3 QXGA @ daylight LED, Core i7 7820HQ, 32 GB RAM, SSD
X62t, SXGA+ @ daylight LED, Core i7 5500U, 16 GB RAM, SSD
T70 WUXGA RGB-LED, Core i5 7300HQ, 32 GB RAM, SSD

PurpleMelbourne
Posts: 26
Joined: Tue Mar 29, 2016 4:19 am
Location: Melbourne, Australia

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#273 Post by PurpleMelbourne » Sun Jul 09, 2017 9:53 pm

Does anyone know what brands of 4GB DIMMs currently available are best for this mod?
I'd love to be able to clock at 440Mhz for added bandwidth. I've currently got a pair of Micron's and a pair of Samsung's, maybe another pair of Hynix as well, though not sure amongst my pile of laptop carnage.

The CNX brand seem to be about 250% the cost of Micron, Samsung, Hynix and others, so it seems best avoided unless you can spare a lot more money for the small gain.

cadillacmike68
ThinkPadder
ThinkPadder
Posts: 1494
Joined: Fri May 27, 2011 9:19 pm
Location: Florida

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#274 Post by cadillacmike68 » Mon Oct 16, 2017 12:08 am

OK, for those of us who ONLY want to undervolt the nvidia GPU a bit to keep it running a bit cooler (and help keep it from failing if it's an older "bad date" nvidia, or even to not overstress a "good date" GPU):

What do we need?

Is the highsun BIOS ONLY for 4:3 T61Ps or will it work for ALL T61s or something in between and what differences?

Does it have SILIC 2.1 for win 7 activation?

Specifically which BIOS from the thinkpadforum-de site would be the one for a 4:3 T61/ T61p (either nvidia GPU) or 15in WS T61/ T61p (either nvidia GPU)??
600 600X
760LD FUBARd
T21 2647 T22 2647 1@ 1GHz SXGA+ 4 more; T23 2647 1@ 1.2GHz SXGA+ 3 more
T30 2366-88U 2GHz; 2366-83U 1.8G; 5@ 2366-LU0/66U; 2367-KU6 FUBARd
T41 T42 T43
T60 T61 8897 2.4GHz SXGA+; 8898 2.4Ghz; 6463 2@ WSXGA+; 7658 2.5GHz; T61p; 6 more T61s
T500 2
T530 W530

dejoebad
Posts: 2
Joined: Fri Nov 03, 2017 12:36 pm
Location: Cilacap, Indonesia

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#275 Post by dejoebad » Fri Nov 03, 2017 10:35 pm

el-sahef wrote:
Wed Jul 05, 2017 2:45 pm
No, not in this case.
I've done running q9100 too on my dell vostro 1014, which is my wife unused laptop. doing only with 5 pins cutted. my questions is. lubuntu only shown 2 cores, where are the other 2 cores? is this because the other 2 cores not running in sync'ed?

I actually bought 2 pcs q9100. 1st is seems had cracked, 2nd is intact (which I cutted the 5 pins). but, strangely, the 1st cpu (that have cracked like when you fail to overclocked and overheated), is working stable (i've tested it for days, tempratures just like usual core2duo). although it shown only 2 cores too, but the cache shown 12mb. while the 2nd, besides only 2 cores shown, the cache shown 6mb only. and my questions are:
1. which one is the right phenomenon actually?
2. is it possible to have underclocked as low as possible to have actual quad core with this laptop?

sorry, if I out of the line, because I need to confirm that this mods is working on vostro 1014, then threat the same on my lenovo t61.

thank you for your help.

dejoebad
Posts: 2
Joined: Fri Nov 03, 2017 12:36 pm
Location: Cilacap, Indonesia

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#276 Post by dejoebad » Fri Nov 03, 2017 10:52 pm

PurpleMelbourne wrote:
Sun Jul 09, 2017 9:53 pm
Does anyone know what brands of 4GB DIMMs currently available are best for this mod?
I'd love to be able to clock at 440Mhz for added bandwidth. I've currently got a pair of Micron's and a pair of Samsung's, maybe another pair of Hynix as well, though not sure amongst my pile of laptop carnage.

The CNX brand seem to be about 250% the cost of Micron, Samsung, Hynix and others, so it seems best avoided unless you can spare a lot more money for the small gain.
according to intel reference for gm45, the ram capacity is related to cpu max capacity can handle. I assume, if you have gm45 + core2quad, then you can have 2 x 4gb installed.
now, I still working in how to enable all 4 cores (now still only 2 cores shown), with underclocked as low as possible for q9100.
I don't really matter about the maximum clock. I only concern if I can run all 4 cores, then I can install 2 x 4gb.

is anyone would mine to respond?

NonesensE
Sophomore Member
Posts: 138
Joined: Mon May 29, 2017 6:28 am
Location: GF/BS/OHA, Germany

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#277 Post by NonesensE » Sat Nov 04, 2017 2:59 am

If 2x4GB RAM is your only concern, you don't need all this. Works with every Dual core, stock or upgraded, too.
Sometimes coming over from the German forum...
X62s 3rd batch, [s]T61[/s] T70 14" 4:3 1st batch

el-sahef
Sophomore Member
Posts: 131
Joined: Sat Mar 17, 2007 6:20 am
Location: Frankfurt, Germany

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#278 Post by el-sahef » Sat Nov 04, 2017 6:03 pm

To activate the two additional CPU cores you need to load a modified APIC table before booting the OS.

The highsun mod BIOS for T61 has this included. Before this BIOS was available I used a different method to load the modified APIC with the GRUB2 bootloader. It is described here in section four: viewtopic.php?f=29&t=110620&start=30#p728624

You might have to adapt it to the vostro though.
T70 15" 4:3 QXGA @ daylight LED, Core i7 7820HQ, 32 GB RAM, SSD
X62t, SXGA+ @ daylight LED, Core i7 5500U, 16 GB RAM, SSD
T70 WUXGA RGB-LED, Core i5 7300HQ, 32 GB RAM, SSD

nightninja96
Sophomore Member
Posts: 146
Joined: Mon Sep 04, 2017 10:06 am
Location: Susquehanna Pa.

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#279 Post by nightninja96 » Wed Nov 08, 2017 6:05 pm

How would I go about soldering the jumper wire to the board?
Thanks

Farba
Posts: 1
Joined: Sat Jun 16, 2018 11:42 am
Location: Zagreb, Croatia

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#280 Post by Farba » Sat Jun 16, 2018 11:56 am

hello!

I managed to install X9100 on T61 4:3 x3100 board.

Image Image Image Image

I am a long IBM fan, i bought my first IBM(T42 15" dothan/sxga+ ips) in high school and i had 10 years, after it died i cracked that lcd while changing a light bulb but i happened to have UXGA screen as a replacemen in storage, for 10-15 years :) . After i saw Frankenpad T61 somewhere on internet i had to have one -> in a month, bought 4:3 board from ebay, took apart some T60 15" and swapped ips and mbo. last 2 years i was using t8300 til someone on my homecountry pc forum said something about 1066fsb cpu in a T61, and that was pretty much that. Now i have spare UXGA ips if this one dies, i need one more mbo just for insurance. I cant use anything else that this ultimate Frankenpad. I am waiting for SSD, T500 cooler and that is pretty much it..

flashed modified BIOS
flashed spd for 533 cl6 ram(x3100 gpu automatically raises it freq. if flashed with lets say 667 spd, the gpu is in stabile...)
soldered that wire for fsb 200
installed cpu and that is that, it is stabile @2.8ghz 1.05v, and 1.163 for 3.2ghz

theterminator93
ThinkPadder
ThinkPadder
Posts: 1450
Joined: Fri Oct 11, 2013 7:45 pm
Location: North Ridgeville, Ohio, United States
Contact:

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#281 Post by theterminator93 » Mon Sep 24, 2018 8:03 am

I've got my T601F which I've already 1066 modded and I've been running a T9900. I got a reasonable deal on a Q9100 and am contemplating whether to install it in that machine or mod another T61 for it. I do a fair bit of work on battery and don't want to run into a scenario where it won't be stable on battery due to drawing too much power.

Anyone with experience in that regard who can offer insight into what I can expect? Otherwise I'll give it a try, and report back...
T480 with T25 keyboard | T25 | W520 i7-2860QM·Quadro 2000m·IPS FHD | T601F T9900·NVS 140m·LED AFFS UXGA
T420 IPS FHD | X220 IPS FHD | T61p·T61·43·42p|X13 Yoga G3·220T·301·41T·24·23·22|G41|A31p·22m|i1200|TransNote
600|770Z|770|760XD|760EL|701C|755C

wujstefan
ThinkPadder
ThinkPadder
Posts: 1343
Joined: Fri Mar 04, 2011 8:38 am
Location: Tarnow, Poland

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#282 Post by wujstefan » Mon Sep 24, 2018 3:36 pm

I did not notice any severe battery life drop when on quadcore. The comparison I made was between X9100 and Q9100, so pretty much the same :)

It does work perfectly stable on battery. Never had any trouble, and the CPU undervolted nicely.
Too many thinkpads not enough time!
(stable under reduction)

pentode
Posts: 12
Joined: Fri Mar 01, 2019 11:29 am
Location: Toulouse, France

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#283 Post by pentode » Fri Mar 01, 2019 11:47 am

My turn to resurrect this great thread.

I've just attempted this mod (quad-core version) on my T61 14" widescreen. I did my homework, took a methodical approach, and thought I had all my bases covered- that is up until I went to do the pin D22 GTLREF connection and realized that my board doesn't have that D22 via. It's just missing. I can try to upload a photo somewhere but I'm guessing the description is enough.

Is this a known issue (I'm guessing people have come across this before)? Does anyone out there have a workaround- another point to supply GTLREF to D22? I took a look at potential bodges and then decided to just appeal to the wisdom of the forum. It doesn't appear that there's any reasonable way to do it unless there's another trace that leads somewhere else.

Assuming there's no solution to that particular problem, any other potential ways to get a quad-core running in that board? What happens if I just drop in the CPU with all other mods and power it up? Just a no-boot situation?

If that's all this board can do, I'll look for a C2D to test what I already did, but I suspect it's OK. I threw a T4400 in that I had lying around from a previous upgrade and it runs much cooler than the T7300 that was in it- Prime95 was hitting 100C and throttling with the T7300- the T4400 runs all day at 62C and is roughly 10% faster. :)

Thanks for any insight.

FryPpy
Senior Member
Senior Member
Posts: 550
Joined: Thu Dec 06, 2012 3:29 pm
Location: Moscow, Russia

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#284 Post by FryPpy » Fri Mar 01, 2019 1:49 pm

pentode wrote:
Fri Mar 01, 2019 11:47 am
My turn to resurrect this great thread.

...I went to do the pin D22 GTLREF connection and realized that my board doesn't have that D22 via. It's just missing.
I have dead T61 14" wide screen mobo and can confirm it have no via for D22.

There is one (theoretical) workaround - interconnecting GTLREF and GTLREF_2 from the socket side. In days of PIN mod, modders use thin wire to interconnect pins (BSEL) in CPU socket.

On centrino (400->533 FSB) photos here
And other 800->1066 FSB mod on T61 here
BAD thing - thin wire have some resistance and i don't know what voltage difference will be between GTLREF and GTLREF_2 with this interconnection.
If you braveheart - think about insulation and path for wire in socket before you begin mod.
pentode wrote:
Fri Mar 01, 2019 11:47 am
If that's all this board can do, I'll look for a C2D to test what I already did, but I suspect it's OK. I threw a T4400 in that I had lying around from a previous upgrade and it runs much cooler than the T7300 that was in it- Prime95 was hitting 100C and throttling with the T7300- the T4400 runs all day at 62C and is roughly 10% faster. :)
Interesting results. T4400 is 800MHz FSB Penryn. So it needs only BIOS mod to start on T61 and works faster that stock T7300.

pentode
Posts: 12
Joined: Fri Mar 01, 2019 11:29 am
Location: Toulouse, France

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#285 Post by pentode » Fri Mar 01, 2019 7:12 pm

Thanks FryPpy. I sat for 5 minutes trying to think of how to do it the "right" way and forgot that "right" is a relative term! :)

I'm going to sleep on this issue and then consider tearing it apart again tomorrow.

For anyone who might be interested, and to respond to your comment, yes, the T4400 (which was in my junk bin) is _always_ faster than a T7300 on this board. I have benchmarks :)

Normally I'd choose the "better" CPU but since I had this one sitting here and already had the machine apart, I tried to be as scientific as possible. I guess the higher clock speed wins vs more cache in this case.

How bad do I want 4 cores?

Thanks again FryPpy!

pentode
Posts: 12
Joined: Fri Mar 01, 2019 11:29 am
Location: Toulouse, France

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#286 Post by pentode » Sun Mar 03, 2019 7:59 am

For anyone with a 14" widescreen T61, I thought I would report back with a few findings that might help anyone interested in attempting the quad-core mod with this model. Maybe it will save someone some time and/or help make the decision about whether to bother.

I was hoping to come back with a success story, but for now, I'll just mention what I've learned.

Following FryPpy's suggestion of using the "pinmod" method I did several experiments on an old socket to see just how feasible this would be in practice. The TL;DR version is:

1) in order to have the CPU sit flush in the socket, your wire must be smaller than .15mm
2) anything thicker than .15mm will not allow the CPU to fit in the socket without putting quite a bit of stress on the pins - enough to bend them (yes, you could enlarge the holes, but see #3 - the wider hole lets things move around in a way that could easily damage the socket)
3) there is a fair risk to the contacts inside the socket, especially with thicker wire. The contacts are very fragile and when you lock the socket into position, depending on the position of your (now 2) wires being pushed back into the contact, you can either bend or break the contacts, or short to an adjacent pin.
4) Working with .1mm magnet wire under these circumstances is a PITA. Thicker wire allows you to form something that will sit in the socket and play nice while you seat the CPU. With .1-.15mm wire, no matter what you do, it will not keep its shape, so it tends to move around while you're placing the CPU. Also, it acts like a spring and flies off at the slightest provocation. A sneeze will definitely disappear your carefully crafted wire :)

So, after having spend the better part of a day experimenting on this, I'm starting to believe that it's not even sensible to attempt if you value your time at all. If the pins were adjacent, it would be so much easier than with them at opposite ends of the socket. I'm certain that it _could_ work as I think .15mm copper is thick enough to not experience any meaningful voltage drop (on .6 volts or whatever is supposed to be there) over 25-30mm.

In a way, I'm happy to take the risk just to see if it will work, but I also hate to potentially kill a laptop that is working fine otherwise. Perhaps I should just stick a fast C2D in it and look for another board to mod for the quad.

pentode
Posts: 12
Joined: Fri Mar 01, 2019 11:29 am
Location: Toulouse, France

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#287 Post by pentode » Sun Mar 03, 2019 2:33 pm

I have successfully booted the Q9000 on the widescreen 14" T61. I'll explain in more detail if anyone out there is interested and wants to attempt this, but for now I have a question.

I see all 4 cores in windows (2 are running at lower speed which I understand) but the CPU is running at 1.5ghz. Should I infer from this that the bus speed mod was therefore unsuccessful? I followed one of the posts in this thread which seemed at the time to be definitive, but I now see that there was a back-and-forth about where to cut the trace on this particular board. The original poster (who seemed successful) cut the trace close to the CPU, but there is a post later on which shows the trace being cut between the SMD component (is that a resistor or a cap? I can't see with my eyes and magnifying glass) and a via directly next to it.

I'm pretty sure it's the same trace (I did check with the meter) but obviously I could be wrong. Assuming it is the same trace, does the location of the cut matter? maybe something on the other end of that via needs the connection to the CPU?

Any insight is appreciated. I'm going to go back and try to find those posts.

edit: here's the critical post viewtopic.php?p=753937#p753937
Last edited by pentode on Sun Mar 03, 2019 5:37 pm, edited 3 times in total.

pentode
Posts: 12
Joined: Fri Mar 01, 2019 11:29 am
Location: Toulouse, France

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#288 Post by pentode » Sun Mar 03, 2019 5:17 pm

Not sure it's proper form to keep adding new posts or whether I should go back and edit previous ones... but for anyone who might stumble across these few posts regarding the 14" widescreen T61, hopefully this info will be helpful.

The machine is now working completely as expected. I went back and fixed the suspect trace and cut it in the spot suggested in a later post (still haven't found it- I've been soldering) and that corrected the bus speed issue. It appears that on other models, that trace goes straight through, but on this model there's a via which takes the signal 2 places (apparently)

If I can find an appropriate place to upload and link photos, I'll compose a how-to for this model in case anyone is interested. I developed a method that should be (relatively) easily repeatable. Rather than attempt to create a magnet-wire link traversing the socket, and due to the fact that there are very slight indentations along the edge of the socket in play, I decided it would be much safer to run the wire from D22 straight off the edge of the socket, tape it to the roll cage, then loop it around, tape at the other side, and into the GTLREF pin. I can go into a lot more detail and I have photos.

I'd be especially curious to know if anyone out there has torn the lid off a socket P socket. I removed the lid from a socket 478 (which is very similar if not pin compatible) and realized that on that socket, each contact is segregated in its own little plastic cell... so much less chance of shorting anything- assuming the socket P is a similar construction. It was that assumption that led me to take the risk :) It'll be amusing to find out I was wrong.

Bottom line: the quad mod is possible on a widescreen 14" x3100 board.

All credit goes to FryPpy for planting the seed and pointing me in the right direction re: pinmod. Huge kudos to the people who created this hack as well. This is what computers are supposed to be about - soldering! :)

Lister
Freshman Member
Posts: 98
Joined: Tue Jul 28, 2015 11:35 am
Location: Czech Republic

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#289 Post by Lister » Mon Mar 04, 2019 1:38 am

Hi,
welcome to the forum. I've read your posts and congratulations to a successful mod. I have a 4:3 Frankenpads, but you can upload photo for other members.
Apple MacBook Pro Retina 15,4 Mid 2015 | i7 4770HQ | | Intel Iris Pro 5200 | 16GB Ram | Samsung ssPolaris 1TB
IBM T60 | UXGA AFFS | T7600 | ATi X1400 | Samsung 3GB | Intel 535 240GB

pentode
Posts: 12
Joined: Fri Mar 01, 2019 11:29 am
Location: Toulouse, France

#290 Post by pentode » Mon Mar 11, 2019 12:28 pm

Has anyone out there managed to get voltage settings and Speedstep working on the quad core mod under linux?

I've got everything configured according to the instructions mentioned in the "master" post on page 2 of this thread. Linux seems to recognize and use the .aml files on boot, yet the system behaves as though no changes have been made. Temps are higher than in windows, speedstep only seems to be working on cores 0 and 1, the voltage on those cores varies (when it should be fixed? - undervolted).

It's funny, I dove into this mod just assuming that the linux end of things would be no problem. Throttlestop under Windows works beautifully and is (dangerously) simple to install and use. I'm almost shocked at the lack of info on the linux side. The c2ctl documentation is incredibly sparse and appears to refer to other documentation that doesn't exist... Dead links everywhere you turn... Even Intel's acpi documentation is byzantine and often broken. Reminds me of linux circa-1996 :)

If anyone out there has any insight on how to make the ACPI/c2ctl/voltage/multiplier stuff actually work properly under linux, I'd be much obliged to hear what you have to say!

Other than this linux snag, the mod is working fantastically with a Q9000. All that I hoped for and more.

p.s. can someone confirm it's ok to post images directly on this forum? It seems to be within the rules, but everyone else posts images elsewhere

ThaGroove
Posts: 1
Joined: Sat Apr 06, 2019 2:48 pm
Location: Kiel, Germany

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#291 Post by ThaGroove » Sat Apr 06, 2019 3:00 pm

Hi,
is here anybody who can tell me, where to get the APIC for Quad Core Support (Socket P for Intel QX 9300, Q 9100, Q 9000) and how to flash it ???
I think the Apic means the drivers for the Bios, right...!?
The actual Bios Mod is installed already...
Its an InsydeH20 Bios, I think Revision 3.5, now without Whitelist and with advanced features...
I think the QX 9300 seems to work partially but with black screen only...
My notebook boots twice and the second time stays on...
Its a different notebook

I know that the Bios of the HP EliteBook 8730w Mobile Workstation must include these files but the Bios is an EXE File :cry:

Screamer
Senior Member
Senior Member
Posts: 585
Joined: Sun Apr 15, 2018 7:15 am

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#292 Post by Screamer » Sun Apr 07, 2019 5:27 am

You could try to search for a BIOS that contains a modified APIC for proper support for the Core 2 Quad processors, in the notebookreview forums.

No, the APIC is just a part of the ACPI subsystem. I don't think it is a 'driver' for the BIOS.

I believe that your current situation is an unstable or a result of poor programming in the BIOS. I doubt it has something to do with the APIC table, since a lot of 8730w dual-to-quad-core users are talking about instability or a problem that is always fully loading 1 core. I might be wrong, but I believe that the problem lies somewhere else in the BIOS.

Sorry, but I don't think I can offer much help regarding this problem. You might have better luck asking for help in the notebookreview forums.

pentode
Posts: 12
Joined: Fri Mar 01, 2019 11:29 am
Location: Toulouse, France

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#293 Post by pentode » Sun Apr 14, 2019 5:54 am

I've set this project aside for a few weeks as, despite the fact that I felt I was getting very close to success, it became too frustrating to continue. In brief, here are the symptoms:

The machine (T61-Q9000) runs flawlessly under Windows. It will run for days on end and is rock solid under all conditions and loads between 0-100%. Under Linux, the machine randomly reboots- the shortest time it lasted was 10 seconds and the longest was 10 hours. It doesn't appear to be load related as I've run Prime95 on it for several hours and it was fine. It typically reboots while I'm not actively using machine- either it is just sitting there, or often it happens playing a Youtube video. I can't replicate the conditions that cause it to reboot- it does seem random.

I have a few ideas, but since the computer runs fine under Windows, I'm going to consider my modified DSDT/SSDT files as suspect until I either get rid of all the errors I'm getting or I understand completely why they're not in play.

I'm hoping that someone out there with a much better understanding of ACPI can take a look at these error messages and point me in the right direction.

Code: Select all

 [    0.000000] BIOS-e820: [mem 0x00000000bf6b0000-0x00000000bf6cbfff] ACPI data
[    0.000000] BIOS-e820: [mem 0x00000000bf6cc000-0x00000000bf6fffff] ACPI NVS
[    0.000000] ACPI: DSDT ACPI table found in initrd [kernel/firmware/acpi/dsdt.aml][0xe4c8]
[    0.000000] ACPI: SSDT ACPI table found in initrd [kernel/firmware/acpi/ssdt1.aml][0x956]
[    0.000000] ACPI: SSDT ACPI table found in initrd [kernel/firmware/acpi/ssdt2.aml][0x262]
[    0.000000] ACPI: SSDT ACPI table found in initrd [kernel/firmware/acpi/ssdt3.aml][0x1ed]
[    0.000000] ACPI: SSDT ACPI table found in initrd [kernel/firmware/acpi/ssdt4.aml][0x1f9]
[    0.000000] ACPI: SSDT ACPI table found in initrd [kernel/firmware/acpi/ssdt5.aml][0x31b]
[    0.000000] ACPI: SSDT ACPI table found in initrd [kernel/firmware/acpi/ssdt6.aml][0x245]
[    0.000000] ACPI: SSDT ACPI table found in initrd [kernel/firmware/acpi/ssdt7.aml][0x116]
[    0.000000] ACPI: SSDT ACPI table found in initrd [kernel/firmware/acpi/ssdt8.aml][0x87d]
[    0.000000] ACPI: SSDT ACPI table found in initrd [kernel/firmware/acpi/ssdt9.aml][0x1ed]
[    0.000000] modified: [mem 0x00000000bf69f000-0x00000000bf6af545] ACPI data
[    0.000000] modified: [mem 0x00000000bf6b0000-0x00000000bf6cbfff] ACPI data
[    0.000000] modified: [mem 0x00000000bf6cc000-0x00000000bf6fffff] ACPI NVS
[    0.000000] ACPI: Early table checksum verification disabled
[    0.000000] ACPI: RSDP 0x00000000000F68C0 000024 (v02 LENOVO)
[    0.000000] ACPI: XSDT 0x00000000BF6BB604 000094 (v01 LENOVO TP-7V    00002270  LTP 00000000)
[    0.000000] ACPI: FACP 0x00000000BF6BB700 0000F4 (v03 LENOVO TP-7V    00002270 LNVO 00000001)
[    0.000000] ACPI BIOS Warning (bug): 32/64X length mismatch in FADT/Gpe0Block: 64/32 (20170831/tbfadt-603)
[    0.000000] ACPI BIOS Warning (bug): Optional FADT field Gpe1Block has valid Address but zero Length: 0x000000000000102C/0x0 (20170831/tbfadt-658)
[    0.000000] ACPI: Table Upgrade: override [DSDT-LENOVO-TP-7V   ]
[    0.000000] ACPI: DSDT 0x00000000BF6BBB1D Physical table override, new table: 0x00000000BF69F000
[    0.000000] ACPI: DSDT 0x00000000BF69F000 00E4C8 (v01 LENOVO TP-7V    00002271 INTL 20180105)
[    0.000000] ACPI: FACS 0x00000000BF6E4000 000040
[    0.000000] ACPI: FACS 0x00000000BF6E4000 000040
[    0.000000] ACPI: Table Upgrade: override [SSDT-LENOVO-TP-7V   ]
[    0.000000] ACPI: SSDT 0x00000000BF6BB8B4 Physical table override, new table: 0x00000000BF6AD4C8
[    0.000000] ACPI: SSDT 0x00000000BF6AD4C8 000956 (v01 LENOVO TP-7V    00002271 INTL 20180105)
[    0.000000] ACPI: ECDT 0x00000000BF6CBB59 000052 (v01 LENOVO TP-7V    00002270 LNVO 00000001)
[    0.000000] ACPI: TCPA 0x00000000BF6CBBAB 000032 (v02 LENOVO TP-7V    00002270 LNVO 00000001)
[    0.000000] ACPI: APIC 0x00000000BF6CBBDD 000078 (v01 LENOVO TP-7V    00002270      00000000)
[    0.000000] ACPI: MCFG 0x00000000BF6CBC55 00003C (v01 LENOVO TP-7V    00002270 LNVO 00000001)
[    0.000000] ACPI: HPET 0x00000000BF6CBC91 000038 (v01 LENOVO TP-7V    00002270 LNVO 00000001)
[    0.000000] ACPI: SLIC 0x00000000BF6CBCC9 000176 (v01 LENOVO TP-7V    00002270  LTP 00000000)
[    0.000000] ACPI: BOOT 0x00000000BF6CBF66 000028 (v01 LENOVO TP-7V    00002270  LTP 00000001)
[    0.000000] ACPI: ASF! 0x00000000BF6CBF8E 000072 (v16 LENOVO TP-7V    00002270 PTL  00000001)
[    0.000000] ACPI: Table Upgrade: override [SSDT-LENOVO-TP-7V   ]
[    0.000000] ACPI: SSDT 0x00000000BF6E2655 Physical table override, new table: 0x00000000BF6ADE1E
[    0.000000] ACPI: SSDT 0x00000000BF6ADE1E 000262 (v01 LENOVO TP-7V    00002271 INTL 20180105)
[    0.000000] ACPI: Table Upgrade: override [SSDT-LENOVO-TP-7V   ]
[    0.000000] ACPI: SSDT 0x00000000BF6E28B4 Physical table override, new table: 0x00000000BF6AE080
[    0.000000] ACPI: SSDT 0x00000000BF6AE080 0001ED (v01 LENOVO TP-7V    00002271 INTL 20180105)
[    0.000000] ACPI: Table Upgrade: override [SSDT-LENOVO-TP-7V   ]
[    0.000000] ACPI: SSDT 0x00000000BF6E295A Physical table override, new table: 0x00000000BF6AE26D
[    0.000000] ACPI: SSDT 0x00000000BF6AE26D 0001F9 (v01 LENOVO TP-7V    00002271 INTL 20180105)
[    0.000000] ACPI: Table Upgrade: override [SSDT-LENOVO-TP-7V   ]
[    0.000000] ACPI: SSDT 0x00000000BF6E2E51 Physical table override, new table: 0x00000000BF6AE466
[    0.000000] ACPI: SSDT 0x00000000BF6AE466 00031B (v01 LENOVO TP-7V    00002271 INTL 20180105)
[    0.000000] ACPI: Table Upgrade: install [SSDT- PmRef- Cpu0Ist]
[    0.000000] ACPI: SSDT 0x00000000BF6AE781 000245 (v01 PmRef  Cpu0Ist  00000101 INTL 20180105)
[    0.000000] ACPI: Table Upgrade: install [SSDT- PmRef- Cpu1Cst]
[    0.000000] ACPI: SSDT 0x00000000BF6AE9C6 000116 (v01 PmRef  Cpu1Cst  00000101 INTL 20180105)
[    0.000000] ACPI: Table Upgrade: install [SSDT- PmRef- Cpu0Cst]
[    0.000000] ACPI: SSDT 0x00000000BF6AEADC 00087D (v01 PmRef  Cpu0Cst  00000101 INTL 20180105)
[    0.000000] ACPI: Table Upgrade: install [SSDT- PmRef- Cpu1Ist]
[    0.000000] ACPI: SSDT 0x00000000BF6AF359 0001ED (v01 PmRef  Cpu1Ist  00000101 INTL 20180105)
[    0.000000] ACPI: Local APIC address 0xfee00000
[    0.000000] ACPI: PM-Timer IO Port: 0x1008
[    0.000000] ACPI: Local APIC address 0xfee00000
[    0.000000] ACPI: LAPIC_NMI (acpi_id[0x00] high edge lint[0x1])
[    0.000000] ACPI: LAPIC_NMI (acpi_id[0x01] high edge lint[0x1])
[    0.000000] ACPI: INT_SRC_OVR (bus 0 bus_irq 0 global_irq 2 dfl dfl)
[    0.000000] ACPI: INT_SRC_OVR (bus 0 bus_irq 9 global_irq 9 high level)
[    0.000000] ACPI: IRQ0 used by override.
[    0.000000] ACPI: IRQ9 used by override.
[    0.000000] Using ACPI (MADT) for SMP configuration information
[    0.000000] ACPI: HPET id: 0x8086a201 base: 0xfed00000
[    0.000000] ACPI: Core revision 20170831
[    0.000000] ACPI: 10 ACPI AML tables successfully acquired and loaded
[    0.064855] PM: Registering ACPI NVS region [mem 0xbf6cc000-0xbf6fffff] (212992 bytes)
[    0.064855] ACPI: bus type PCI registered
[    0.064855] acpiphp: ACPI Hot Plug PCI Controller Driver version: 0.5
[    0.068291] ACPI: Added _OSI(Module Device)
[    0.068291] ACPI: Added _OSI(Processor Device)
[    0.068291] ACPI: Added _OSI(3.0 _SCP Extensions)
[    0.068291] ACPI: Added _OSI(Processor Aggregator Device)
[    0.068291] ACPI: Added _OSI(Linux-Dell-Video)
[    0.068291] ACPI: Added _OSI(Linux-Lenovo-NV-HDMI-Audio)
[    0.068291] ACPI: Added _OSI(Linux-HPI-Hybrid-Graphics)
[    0.068291] ACPI: EC: EC started
[    0.068291] ACPI: EC: interrupt blocked
[    0.069677] ACPI: \: Used as first EC
[    0.069680] ACPI: \: GPE=0x12, EC_CMD/EC_SC=0x66, EC_DATA=0x62
[    0.069682] ACPI: \: Used as boot ECDT EC to handle transactions
[    0.069903] ACPI: Executed 10 blocks of module-level executable AML code
[    0.076009] ACPI: [Firmware Bug]: BIOS _OSI(Linux) query ignored
[    0.093037] ACPI: Dynamic OEM Table Load:
[    0.093053] ACPI: SSDT 0xFFFF8FB1B6718800 000240 (v01 PmRef  Cpu0Ist  00000100 INTL 20050513)
[    0.093074] ACPI Error: [_PPC] Namespace lookup failure, AE_ALREADY_EXISTS (20170831/dswload-378)
[    0.093085] ACPI Exception: AE_ALREADY_EXISTS, During name lookup/catalog (20170831/psobject-252)
[    0.093553] ACPI: Dynamic OEM Table Load:
[    0.093568] ACPI: SSDT 0xFFFF8FB1B65B7000 00085E (v01 PmRef  Cpu0Cst  00000100 INTL 20050513)
[    0.093585] ACPI Error: [CMW1] Namespace lookup failure, AE_ALREADY_EXISTS (20170831/dswload-378)
[    0.093593] ACPI Exception: AE_ALREADY_EXISTS, During name lookup/catalog (20170831/psobject-252)
[    0.093680] ACPI: Dynamic OEM Table Load:
[    0.093680] ACPI: SSDT 0xFFFF8FB1B7148B00 0000C8 (v01 PmRef  Cpu1Ist  00000100 INTL 20050513)
[    0.093680] ACPI Error: [_PPC] Namespace lookup failure, AE_ALREADY_EXISTS (20170831/dswload-378)
[    0.093680] ACPI Exception: AE_ALREADY_EXISTS, During name lookup/catalog (20170831/psobject-252)
[    0.093680] ACPI: Dynamic OEM Table Load:
[    0.093680] ACPI: SSDT 0xFFFF8FB1B6595CC0 000085 (v01 PmRef  Cpu1Cst  00000100 INTL 20050513)
[    0.093680] ACPI Error: [_CST] Namespace lookup failure, AE_ALREADY_EXISTS (20170831/dswload-378)
[    0.093680] ACPI Exception: AE_ALREADY_EXISTS, During name lookup/catalog (20170831/psobject-252)
[    0.093680] ACPI: Interpreter enabled
[    0.093680] ACPI: (supports S0 S3 S4 S5)
[    0.093680] ACPI: Using IOAPIC for interrupt routing
[    0.093680] PCI: Using host bridge windows from ACPI; if necessary, use "pci=nocrs" and report a bug
[    0.093889] ACPI: Enabled 5 GPEs in block 00 to 1F
[    0.108754] ACPI: Power Resource [PUBS] (on)
[    0.110336] acpi PNP0C0A:01: ACPI dock station (docks/bays count: 1)
[    0.113000] acpi LNXIOBAY:00: ACPI dock station (docks/bays count: 2)
[    0.123397] acpi IBM0079:00: ACPI dock station (docks/bays count: 3)
[    0.124475] ACPI: PCI Interrupt Link [LNKA] (IRQs 3 4 5 6 7 9 *10 11)
[    0.124740] ACPI: PCI Interrupt Link [LNKB] (IRQs 3 4 5 6 7 9 10 *11)
[    0.125003] ACPI: PCI Interrupt Link [LNKC] (IRQs 3 4 5 6 7 9 10 *11)
[    0.125265] ACPI: PCI Interrupt Link [LNKD] (IRQs 3 4 5 6 7 9 10 *11)
[    0.125528] ACPI: PCI Interrupt Link [LNKE] (IRQs 3 4 5 6 7 9 10 *11)
[    0.125791] ACPI: PCI Interrupt Link [LNKF] (IRQs 3 4 5 6 7 9 10 *11)
[    0.126055] ACPI: PCI Interrupt Link [LNKG] (IRQs 3 4 5 6 7 9 10 *11)
[    0.126318] ACPI: PCI Interrupt Link [LNKH] (IRQs 3 4 5 6 7 9 10 *11)
[    0.126517] ACPI: PCI Root Bridge [PCI0] (domain 0000 [bus 00-ff])
[    0.135372] pci 0000:00:1f.0: quirk: [io  0x1000-0x107f] claimed by ICH6 ACPI/GPIO/TCO
[    0.144005] ACPI: EC: interrupt unblocked
[    0.144005] ACPI: EC: event unblocked
[    0.144005] ACPI: \_SB_.PCI0.LPC_.EC__: GPE=0x12, EC_CMD/EC_SC=0x66, EC_DATA=0x62
[    0.144005] ACPI: \_SB_.PCI0.LPC_.EC__: Used as boot DSDT EC to handle transactions and events
[    0.144058] ACPI: bus type USB registered
[    0.144203] PCI: Using ACPI for IRQ routing
[    0.177796] pnp: PnP ACPI init
[    0.179072] system 00:00: Plug and Play ACPI device, IDs PNP0c01 (active)
[    0.179436] system 00:01: Plug and Play ACPI device, IDs PNP0c02 (active)
[    0.179576] pnp 00:02: Plug and Play ACPI device, IDs PNP0b00 (active)
[    0.179623] pnp 00:03: Plug and Play ACPI device, IDs PNP0303 (active)
[    0.179677] pnp 00:04: Plug and Play ACPI device, IDs IBM0057 PNP0f13 (active)
[    0.181562] pnp: PnP ACPI: found 5 devices
[    1.149744] ACPI: AC Adapter [AC] (on-line)
[    1.149883] ACPI: Lid Switch [LID]
[    1.149977] ACPI: Sleep Button [SLPB]
[    1.150083] ACPI: Power Button [PWRF]
[    1.157495] ACPI: Thermal Zone [THM0] (29 C)
[    1.159655] ACPI: Thermal Zone [THM1] (31 C)
[    1.177137] ACPI: Battery Slot [BAT0] (battery present)
[    2.231884] ata3.00: ACPI cmd ef/02:00:00:00:00:a0 (SET FEATURES) succeeded
[    2.231887] ata3.00: ACPI cmd f5/00:00:00:00:00:a0 (SECURITY FREEZE LOCK) filtered out
[    2.231890] ata3.00: ACPI cmd ef/10:03:00:00:00:a0 (SET FEATURES) filtered out
[    2.235312] ata3.00: ACPI cmd ef/02:00:00:00:00:a0 (SET FEATURES) succeeded
[    2.235315] ata3.00: ACPI cmd f5/00:00:00:00:00:a0 (SECURITY FREEZE LOCK) filtered out
[    2.235317] ata3.00: ACPI cmd ef/10:03:00:00:00:a0 (SET FEATURES) filtered out
[    4.025134] thinkpad_acpi: ThinkPad ACPI Extras v0.25
[    4.025440] ACPI Warning: SystemIO range 0x0000000000001028-0x000000000000102F conflicts with OpRegion 0x0000000000001000-0x000000000000107F (\_SB.PCI0.LPC.PMIO) (20170831/utaddress-247)
[    4.025447] ACPI: If an ACPI driver is available for this device, you should use it instead of the native driver
[    4.025451] ACPI Warning: SystemIO range 0x00000000000011B0-0x00000000000011BF conflicts with OpRegion 0x0000000000001180-0x00000000000011BF (\_SB.PCI0.LPC.LPIO) (20170831/utaddress-247)
[    4.025456] ACPI: If an ACPI driver is available for this device, you should use it instead of the native driver
[    4.025458] ACPI Warning: SystemIO range 0x0000000000001180-0x00000000000011AF conflicts with OpRegion 0x0000000000001180-0x00000000000011BF (\_SB.PCI0.LPC.LPIO) (20170831/utaddress-247)
[    4.025463] ACPI: If an ACPI driver is available for this device, you should use it instead of the native driver
[    4.068475] thinkpad_acpi: ACPI backlight control delay disabled
[    4.069176] thinkpad_acpi: This ThinkPad has standard ACPI backlight brightness control, supported by the ACPI video driver
[    4.080041] thinkpad_acpi: Standard ACPI backlight interface available, not loading native one
[    4.344252] ACPI: Video Device [VID] (multi-head: yes  rom: no  post: no)

As you can see, all 10 of my modified tables (actually 1 or 2 of them stayed stock but I made a full set for reasons I won't go into now) are apparently successfully loaded _and_ overriding the hardware tables. However, for reasons that completely escape me, 4 of the BIOS tables are attempting to load after my set as _dynamic_ tables. These are the CPU0CST/CPU0IST and CPU1CST/CPU1IST tables. I don't know why they're being loaded, but it seems to me that if any of those variables related to c-states and power states are screwed up, that might account for my "random" reboots.

The second obvious problem is the address conflicts. It appears they're related to the backlight, but I honestly have no idea. It seems logical, though, that if the same memory address is being used by 2 devices, this could cause a reboot.

Does anyone out there know why those 4 tables are being loaded dynamically after my modified tables seem to load properly? And why just those 4? It seems suspicious since those are specifically CPU-power related. Is there a way to prevent dynamic tables from being loaded? I can't find any ACPI kernel option that can do it. Does anyone know if either of these problems are likely to be the source of my problem, or if I'm looking in the wrong direction altogether?

Any and all insight is very much appreciated. I'd love to get this machine running properly under Linux.

p.s. for anyone attempting this in 2019, one step that I don't believe is mentioned in the how-to posts is that as of now, the tables will not overwrite the old BIOS tables unless you bump the revision number. If you don't change the revision number, they are apparently loaded but unused as the kernel either uses the newest table version, or the first one it loads.
Last edited by pentode on Sun Apr 14, 2019 6:09 am, edited 1 time in total.

pentode
Posts: 12
Joined: Fri Mar 01, 2019 11:29 am
Location: Toulouse, France

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#294 Post by pentode » Sun Apr 14, 2019 6:00 am

ThaGroove wrote:
Sat Apr 06, 2019 3:00 pm
Hi,
is here anybody who can tell me, where to get the APIC for Quad Core Support (Socket P for Intel QX 9300, Q 9100, Q 9000) and how to flash it ???
I think the Apic means the drivers for the Bios, right...!?
The actual Bios Mod is installed already...
Its an InsydeH20 Bios, I think Revision 3.5, now without Whitelist and with advanced features...
I think the QX 9300 seems to work partially but with black screen only...
My notebook boots twice and the second time stays on...
Its a different notebook

I know that the Bios of the HP EliteBook 8730w Mobile Workstation must include these files but the Bios is an EXE File :cry:
I want to make sure I understand before I try to answer- this is a T61 that you've done the quad-core mod on? I can't say for sure (and I'm not familiar with the bios you mentioned, but the only bios I'm sure will boot these quad cores properly is one of the Chinese files from the how-to on the German site. If you give me more information, I may be able to steer you in the right direction, but I can tell you for sure that these machines will not boot without properly-sorted-out ACPI tables. Black screen is one symptom I've seen.

pentode
Posts: 12
Joined: Fri Mar 01, 2019 11:29 am
Location: Toulouse, France

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#295 Post by pentode » Thu May 02, 2019 9:52 am

In case anyone trying to do the quad-core mod and is having problems under linux, I thought I would share my latest findings on the problems I've been having. Short catch-up: while my T61 with Q9000 runs just fine under windows, I've been having a random reboot problem under linux, which I attributed to problems with ACPI files- either compatibility issues with the original files posted by el-sahef or just plain errors in my own versions. I'm not an expert by any means in ACPI, nor in Intel CPU architecture, nor in linux kernel hacking, but my problem is beginning to look a bit more subtle than I first thought.

In summary, here's what I have found:

1) Running Ubuntu 18.04, with or without modified ACPI files, the machine freezes at totally random intervals- when playing sound, it will loop a half-second sample of whatever was in the buffer for about 5 seconds, then reboot

2) Under Debian Stretch, with or without modified ACPI, the machine exhibits a similar behavior where it stutters, sometimes looping whatever sound was playing for a couple seconds, but then recovers and shows no apparent ill effects. This happens roughly every minute or so.

3) Under Slackware and Devuan (both non-systemd coincidentally and with stock ACPI) the machine runs perfectly normally. I've run it for 48+ hours under each of these distributions with no apparently symptoms.

4) Under Ubuntu, whenever the machine reboots, the last thing shown in the logs is always a warning message about "perf: interrupt took too long" Under Debian, while the machine is stuttering, dmesg reports the same "perf: interrupt took too long" warning happening at the same time. There are other errors which seem to indicate that somehow, one of the high performance clocks is developing a bigger and bigger skew. Probably not coincidentally, I've learned that perf is implicated whenever the CPU is scaling.

So, my shoot-from-the-hip semi-scientific deduction at this point is that somehow, with certain kernels, under certain circumstances, the timing between cores 0/1 and 2/3 is somehow becoming destabilized to the point where the kernel gets really unhappy. Some kernels in some distros seem to weather this without a problem and some don't.

If you've read my previous posts, I think the information about ACPI tables not matching and not being loaded as expected is still a valid concern, but probably not the cause of my issues.

If anyone out there has any theories as to what might be happening and how to prevent it, I'd love to hear them. In the meantime, I'll just keep chipping away in my spare time. At least I have the machine running stably under Linux, which is all I wanted in the first place, albeit with a non-systemd distro, which I was considering doing anyway since I'm old and grumpy. :)

atagunov
Senior Member
Senior Member
Posts: 921
Joined: Thu Apr 02, 2020 3:11 pm
Location: London, UK

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#296 Post by atagunov » Sun Apr 26, 2020 8:19 pm

pentode wrote:
Thu May 02, 2019 9:52 am
If anyone out there has any theories as to what might be happening and how to prevent it, I'd love to hear them
Pentode, I feel your pain
QX9100 Ubuntu T60... :bouncing-bird: :bouncing-bird: :bouncing-bird:

Is this hope?
recommended method

In the meantime, a whole host of Mod BIOS versions of highsun are available on 51nb.com, including those specially for Core2 quads that already contain the modified APIC table, so that apart from the BIOS flash, no further software changes are required
Files are here. Downloadable.

P.S. buying 14" 4:3 T61 mobo, CPU, fan, invertor, leds, RAM in the UK/EU.. Pls PM :)
X220, 2 *T520

atagunov
Senior Member
Senior Member
Posts: 921
Joined: Thu Apr 02, 2020 3:11 pm
Location: London, UK

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#297 Post by atagunov » Mon Apr 27, 2020 2:31 pm

@pentode, while I still don't have my Frankenpad hardware I have read up as much as I could on the project. Here are my thoughts on your issues. These issues may soon be mine so I've given them a lot of thought.

It seems you're
- using some BIOS other than Highsun's
- loading ACPI (DDST+SSDT) tables which are a mix of tables you extracted from your running system (this should in fact be a replica of what you have in your BIOS) with tables posted online earlier elsewhere

As my goal is to run Ubuntu here's what I think could be done to improve your Ubuntu situation:

- try 'acpi_no_auto_ssdt' kernel parameter via grub (somehow)
- use 'initrd' method of packaging DDST/SSDT
- use Highsun's BIOS (my prev. post) and do not load any tables of your won - rely on those packaged in Highsun's BIOS
- possibly pass 'acpi_no_auto_ssdt' to kernel + only use those tables published elsewhere and forget about your own
- choose whichever BIOS looks best to you, I'd suggest Higsun's, dump tables from a system running with that BIOS and very judicially add/modify extra 2 CPU-s as described on wayback machine c2ctl page (btw it explains in a rather nice and clear language how to first try things with 'c2ctl' and then encode the resulting found voltage/freq pairs into SSDT); only use tables published elsewhere as a reference; this looks like a difficult path because it seems you'd have to fix all compilation errors in the DDST/SSDT-s however many there may be
- it seems on top of that one may still need to use 'c2ctl' to enable speedstep (to be checked if that's really needed)

I'm sorry if you have already tried many of this of if nothing of this helps; I will probably go on to do the same this year once I have slowly acquired the necessary hardware. My Plan B if all this fails is just to run with X9100 instead of QX9100 :)
X220, 2 *T520

Sunshine701c
Posts: 36
Joined: Sat Mar 14, 2020 6:57 pm
Location: Portland, OR

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#298 Post by Sunshine701c » Tue Nov 09, 2021 3:25 am

I figured I'd post just to bump this amazingly helpful thread. I just finished this mod on my 14" T61 mobo, currently posting from it with a T9400 installed, in a 15" case, 1600x1200 LED LCD. It has NVIDIA graphics, and while I didn;t see anyone posting specifically about that board, I was able to do the mod successfully, if anyone needs help finding the right traces to cut and solder for this motherboard specifically I think I might be able to help.
My only other concern is that I want to get a q9000 or q9100 in this thing. I already did the 1 wire mod where I connected the 2 on-board resistors to the CPU pin, is there anything else that's needed for the Quad-Core mod, other than cutting the pins on the CPU itself?
Also, how is the performance of the q9000 vs the T9400? Will 2 more cores at a lower frequency speed up anything significantly on this laptop?

kfzhu1229
Senior ThinkPadder
Senior ThinkPadder
Posts: 2484
Joined: Tue Sep 22, 2015 10:59 pm
Location: Toronto, Canada

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#299 Post by kfzhu1229 » Sat Nov 13, 2021 7:01 pm

I end up referencing to this tutorial to make use of a P8700 that I have lying around but all my remaining Core 2 duo laptops don't support 1066mhz FSB.
I have two candidates for the P8700, a HP G72 with the low end version of the motherboard that has the GL40 chipset (a "disabled" GM45 chipset that runs slower and doesn't support 1066mhz FSB), and a Dell XPS M1330 with the 965GM chipset that we all know of.
I first tried it on the GL40 chipset, after much trial and error, unfortunately it seems like the GL40 chipset was "disabled" from a GM45 chipset for a reason - bad bins. When I first performed the 1066mhz -> 800Mhz mod by adding a jumper wire from BSEL1 to VCC on the socket, the computer worked then, but at the step of isolating BSEL1 and then grounding that on the clock generator (which is quite easy on this thanks to HP using resistors that I can simply remove rather than cutting a trace), no matter what I do and what RAM I try, before and after SPD flashings, it simply won't POST. My interpretation is probably that the GL40 chipset is simply not stable with a clock of 266mhz over 200mhz.
Obviously I can't just let a 17" laptop have a Core 2 duo that essentially runs at only 1.9Ghz, so I have to abort that and move on to my next candidate.
The M1330 with the 965GM chipset is pretty standard affair compared to a T61 with integrated graphics. With the same jumper wire mod, the computer worked with the P8700 at the reduced 800Mhz FSB perfectly fine, no BIOS mod needed besides BIOS revision required to support Penryn.
This one I have to unfortunately painstakenly cut off a trace on the motherboard and it's a super short trace for BSEL1 for the clock generator before going to a via. Unfortunately the clock generator is of a QFN package so I can't just lift off a leg or something.
Now I have some questions regarding with SPD modding the RAM. Is it better to mod the RAM with more loose timings or is it better to mod the RAM to operate at 533Mhz and get an effective clock of 710Mhz?
And does that mean if the T9400, T9550, T9600, T9800, T9900 or P8xxx/P9xxx are what will be installed after this mod, no BIOS mod is required on a native Penryn T61?
EDIT: Also all it took for me to get a pair of my Micron 4GB sticks from bootlooping at POST to fully working and passing memtest86 after this mod was literally changing CL timing from 5 to 6! Where the RAM stick usually runs 667mhz with 5-5-5-15 timings, now it's perfectly stable with 886Mhz with 6-5-5-15. Why is the CL timing all that mattered here?
Dell Lat CP MMX-233 64mb 40gb W2k
600 PII-266 416mb 40gb WXP
T23 PIII 1.13ghz 1gb W7
Precision M4300 X9000 8gb 160gb WUXGA Ultrasharp fp W10
T530i 15.6" i7 16gb fp W10
UXGA:
A30p PIII 1.2 1gb W7 (IDTech)
T43p 2.26 2gb fp W10 (Sharp)
Lat C840 P4-2.5 2gb 60gb W7 (Ultrasharp)

kfzhu1229
Senior ThinkPadder
Senior ThinkPadder
Posts: 2484
Joined: Tue Sep 22, 2015 10:59 pm
Location: Toronto, Canada

Re: Install a 1066MHz FSB CPU on T61 (including Core 2 Quad)

#300 Post by kfzhu1229 » Sat Feb 05, 2022 5:52 pm

Well I have been keep coming back to this post for reference for me to do this mod on other laptops running the 965 platform. Latest labrat is my Compaq 6910p and that's successfully modded to 1066Mhz FSB as well (but on this model, sleep no longer works properly, but hibernate/fast startup does)
As for the post above viewtopic.php?p=728624#p728624
And the picture http://abload.de/img/266-fsbe2cow.jpg
Is there no one wondering why you need to cut that wire? A safer approach would be to remove that resistor, turn it around 180 degrees to the bottom, and then solder the free end of the resistor to the capacitor instead? That's what I did in equivalent on my 6910p
Or if you have clumsy hands and bad vision, you could remove that resistor (1KOhm) entirely and solder a much bigger thru hole 1Kohm resistor in its place, this time from the bottom pad to that Vcore capacitor, and I imagine this is much easier to beginners!
And yeah anyways, thanks so much to the people that have figured this out!
Dell Lat CP MMX-233 64mb 40gb W2k
600 PII-266 416mb 40gb WXP
T23 PIII 1.13ghz 1gb W7
Precision M4300 X9000 8gb 160gb WUXGA Ultrasharp fp W10
T530i 15.6" i7 16gb fp W10
UXGA:
A30p PIII 1.2 1gb W7 (IDTech)
T43p 2.26 2gb fp W10 (Sharp)
Lat C840 P4-2.5 2gb 60gb W7 (Ultrasharp)

Post Reply
  • Similar Topics
    Replies
    Views
    Last post

Return to “ThinkPad T60/T61 Series”

Who is online

Users browsing this forum: No registered users and 47 guests