Page 1 of 1
Thinkpad T60 Self-Upgrade @Core2 Duo T7200 + 4GB RAM + 7K100
Posted: Tue Jan 02, 2007 3:07 pm
by Austin_Goh
My Thinkpad T60 has gone through a series of major upgrades to enhance it's performance.
The Model I bought is 2007-4JA (I live in Malaysia), which came with:
- Core Duo T2400 1.83GHz
- 1GB DDR2 667
- Hitachi 60GB 5400rpm
Snapshots when I bought it during October 2006. (Click thumbnails to enlarge images)
img]http://img9.picsplace.to/img9/21/thumbs/IMG_0921.JPG[/img]
img]http://img9.picsplace.to/img9/21/thumbs/IMG_0922.JPG[/img]
img]http://img9.picsplace.to/img9/21/thumbs/IMG_0923.JPG[/img]
img]http://img9.picsplace.to/img9/21/thumbs/IMG_0924.JPG[/img]
Here's the upgrade list:
- Core 2 Duo T7200 2GHz 4MB Cache
- 2x 2GB DDR2 533MHz CL4-4-4-12
- Hitachi Travelstar 7K100 60GB 7200rpm 8MB Buffer SATA
To reduce my cost, I bought a refurbished Dell laptop and swap out 2 components, namely the processor and Hard Disk.
Dell Inspiron 6400 been dismantled:
http://kcbox.net/my/T60_Upgrade/IMG_1015.JPG
http://kcbox.net/my/T60_Upgrade/IMG_0993.JPG
http://kcbox.net/my/T60_Upgrade/IMG_0994.JPG
Lenovo Thinkpad T60 been dismantled:
http://img6.picsplace.to/img6/22/IMG_0987.JPG
http://img6.picsplace.to/img6/22/IMG_0988.JPG
Core 2 Duo T7200 2000MHz 4MB Cache 667FSB EM64T mPGA479 Processor
http://img10.picsplace.to/18/IMG_0998_resize.JPG
T2400 1833MHz 2MB Cache at right hand side
http://img10.picsplace.to/18/IMG_0997_resize.JPG
Hitachi 7K100 60GB 7200rpm 8MB Buffer SATA Hard Disk
http://img10.picsplace.to/18/IMG_1000_resize.JPG
http://img10.picsplace.to/18/IMG_0999_resize.JPG
2pcs Infineon 2GB = 4GB DDR2 533MHz CL4-4-4-12 running dual channel
http://img6.picsplace.to/img6/22/IMG_0981.jpg
http://img6.picsplace.to/img6/22/IMG_0980.jpg
http://img6.picsplace.to/img6/22/IMG_0979.jpg
----------------------------------------------------------------------------------------------------------
Benchmark Results (Click thumbnails to enlarge images)
Super Pi 1MB Calculation: 27seconds
CPU Mark 99: 276points
3D MARK 2001SE 9900 points
3D MARK 03 3554 points
3D MARK 05 1614 points
img]http://img6.picsplace.to/img6/22/thumbs/superpi.JPG[/img]
img]http://img10.picsplace.to/18/thumbs/hdtach.JPG[/img]
img]http://img10.picsplace.to/18/thumbs/ssmm.JPG[/img]
img]http://img8.picsplace.to/img8/23/thumbs/3dmark.JPG[/img]
img]http://img8.picsplace.to/img8/23/thumbs/3dmark05.JPG[/img]
1GB of RAM space will be located for RAM Disk Utility, all Windows system Page File, Temporary Internet Folder, frequently used programs will be here to boost the performance.
HD Speed Benchmark:
Hitachi 7K100 60GB ~77MB/s
RAM Drive 1GB ~3700MB/s
img]http://kcbox.net/my/T60_Upgrade/set_ramdrive_resize.JPG[/img]
img]http://kcbox.net/my/T60_Upgrade/perform ... resize.JPG[/img]
http://kcbox.net/my/R150%20Repainted/IMG_1134.JPG
Now I am satisfied with it, the upgrade progress is not that difficult, must read the hardware service manual first to avoid any physical damage if you want to do so.
TQ.
Click Here to Download Thinkpad T60 Hardware Maintenance Manual
Note from Moderator: No warning in the subject line = no images. Links only.
Posted: Tue Jan 02, 2007 3:42 pm
by rscosworth
nice, very nice
can you tell me if the core2duo desktop cpu's are different from the core2duo laptop cpus?
cheers
Posted: Tue Jan 02, 2007 3:47 pm
by z_24
hmmm..I never know you could upgarde a processor in a TP. Interesting.
Posted: Tue Jan 02, 2007 3:49 pm
by own6volvos
Nice job. Did you notice any problems removing the heatsinks without leaving marks on the screws? Also, did you notice any temperature difference (drop) after putting on some artic silver or better heatsink compound over the lenovo stuff from the factory?
Posted: Tue Jan 02, 2007 4:04 pm
by Austin_Goh
own6volvos wrote:Nice job. Did you notice any problems removing the heatsinks without leaving marks on the screws? Also, did you notice any temperature difference (drop) after putting on some artic silver or better heatsink compound over the lenovo stuff from the factory?
Didn't see any mark on the heatsink screws.
I didn't put any better thermal compound, temperature is around 50 degree celcius when idle, slightly over 70c when full load with 2x SP2004 Prime running.
T2400 has about the same temperature. Room temperature is 27 to 30c.

Posted: Tue Jan 02, 2007 4:13 pm
by Rwang
How much performance saving are you getting out of the Qsoft RAM?
Nice job by the way.
Posted: Tue Jan 02, 2007 4:13 pm
by RonS
Those are really nice pictures, but I have to wonder if it was really worth it. Why not just buy a stock T60p? I have an unaltered T60p, 2007-93U. Comparing with the numbers for your modded system,
I get 5MB/sec faster HD numbers (HDTach)
CPUmark 99 speeds are the same (276)
3DMark03, I got 5064 (vs 3554)
3DMark05, I got 3877 (vs 1614)
All using stock drivers.
Posted: Tue Jan 02, 2007 4:39 pm
by pianowizard
Austin_Goh, did you know that you're slowing down your machine by downgrading from PC2-5300 to PC2-4200? You should have gotten PC2-5300 modules.
Posted: Tue Jan 02, 2007 4:43 pm
by Austin_Goh
RonS wrote:Those are really nice pictures, but I have to wonder if it was really worth it. Why not just buy a stock T60p? I have an unaltered T60p, 2007-93U. Comparing with the numbers for your modded system,
I get 5MB/sec faster HD numbers (HDTach)
CPUmark 99 speeds are the same (276)
3DMark03, I got 5064 (vs 3554)
3DMark05, I got 3877 (vs 1614)
All using stock drivers.
Sigh in Malaysia, the "highest" spec available is merely a T5600 c2d, how I wish I could purchase a T60p with stock config as yours.
Anyway the 3D performance is not necessary as I own a gaming rig already.
I tested all XP 32bit & x64 edition, server 2003 Enterprise Edition, Vista Ultimate 32bit, the available RAM amounted to 3GB only, PAE was enabled though. Any idea how to fully utilize 4GB of RAM?
Posted: Tue Jan 02, 2007 5:17 pm
by Austin_Goh
pianowizard wrote:Austin_Goh, did you know that you're slowing down your machine by downgrading from PC2-5300 to PC2-4200? You should have gotten PC2-5300 modules.
Since the system bus is limited to 667MHz only, by getting 2pcs DDR2 533MHz running @ Dual Channel will be slightly faster than 667MHz Dual Channel.
CAS Latency on 533MHz is CL4, on 667MHz is CL5.
btw if Singlw Channel only, 667MHz will be faster instead.

Posted: Tue Jan 02, 2007 5:30 pm
by pianowizard
Austin_Goh wrote:Since the system bus is limited to 667MHz only, by getting 2pcs DDR2 533MHz running @ Dual Channel will be slightly faster than 667MHz Dual Channel.
I can't follow your rationale here. Because the system bus can support up to 667MHz, you should take advantage of it and use 667MHz memory. 533Mhz is slower than 667MHz. For frequency, the higher the better.
Austin_Goh wrote:CAS Latency on 533MHz is CL4, on 667MHz is CL5.
We had a similar discussion a few weeks back. 533MHz with CL4 is not faster than 667MHz with CL5. This same thread covers other issues as well:
http://forum.thinkpads.com/viewtopic.php?t=24617
BTW, the advantage of dual channel is only about 10% at most.
Posted: Tue Jan 02, 2007 5:36 pm
by Austin_Goh
pianowizard wrote:Austin_Goh wrote:Since the system bus is limited to 667MHz only, by getting 2pcs DDR2 533MHz running @ Dual Channel will be slightly faster than 667MHz Dual Channel.
I can't follow your rationale here. Because the system bus can support up to 667MHz, you should take advantage of it and use 667MHz memory. 533Mhz is slower than 667MHz. For frequency, the higher the better.
Austin_Goh wrote:CAS Latency on 533MHz is CL4, on 667MHz is CL5.
We had a similar discussion a few weeks back. 533MHz with CL4 is not faster than 667MHz with CL5. This same thread covers other issues as well:
http://forum.thinkpads.com/viewtopic.php?t=24617
BTW, the advantage of dual channel is only about 10% at most.
LOL anyway I am willing to use 4GB of 667MHz DDR2 if they are priced the same as 533MHz. Seriously if we are talking about merely 2x 1GB, PC5300 DDR2 667MHz will be my first choice too

Posted: Wed Jan 03, 2007 3:12 am
by own6volvos
Austin_Goh wrote:own6volvos wrote:Nice job. Did you notice any problems removing the heatsinks without leaving marks on the screws? Also, did you notice any temperature difference (drop) after putting on some artic silver or better heatsink compound over the lenovo stuff from the factory?
Didn't see any mark on the heatsink screws.
I didn't put any better thermal compound, temperature is around 50 degree celcius when idle, slightly over 70c when full load with 2x SP2004 Prime running.
T2400 has about the same temperature. Room temperature is 27 to 30c.

Your T7200 temps are pretty high. What exactly are you using to note tempurature of the CPU. Are you using the cpu socket temp provided by the tpfancontrol software, or a program that reads core temps like speedfan or everest?
At roughly 65-70F my T7400 idles around 43-45C under normal use. That is stock untouched heatsinks. Perhaps it might be wise to clean off the heatsinks in your machine with some alcohol, and regrease (thermal compound) them?
Posted: Wed Jan 03, 2007 10:34 am
by Austin_Goh
own6volvos wrote:Austin_Goh wrote:
Didn't see any mark on the heatsink screws.
I didn't put any better thermal compound, temperature is around 50 degree celcius when idle, slightly over 70c when full load with 2x SP2004 Prime running.
T2400 has about the same temperature. Room temperature is 27 to 30c.

Your T7200 temps are pretty high. What exactly are you using to note tempurature of the CPU. Are you using the cpu socket temp provided by the tpfancontrol software, or a program that reads core temps like speedfan or everest?
At roughly 65-70F my T7400 idles around 43-45C under normal use. That is stock untouched heatsinks. Perhaps it might be wise to clean off the heatsinks in your machine with some alcohol, and regrease (thermal compound) them?
should have no difference if under the same ambient temperature as yours.
My ambient temperature is quite high most of the time, if not mistaken during I tested all these benchmarks, ambient temp is 33c.
Another thing is, I used core temp utility to read the actual cpu temp from the integrated thermal probe of the processor.
Now the temperature reading when idling is 46c, night time ambient temp is 27c.
Posted: Thu Jan 04, 2007 2:28 am
by kulivontot
For most applications, the memory bus is rarely starved for bandwidth, thus DDR2-533 vs. DDR2-667 is generally negligible for the same latency values. However, DDR2-533 modules currently are capable of lower latency values than DDR2-667 modules, which does result in a tangible difference in performance. Thus I can understand the rationale in purchasing a high-performance low-latency DDR2-533 module over a moderate-performance higher-latency DDR2-667 module for the same price. There are more factors in raw performance than pure Mhz in most cases.
Posted: Thu Jan 04, 2007 7:01 am
by pianowizard
kulivontot wrote:However, DDR2-533 modules currently are capable of lower latency values than DDR2-667 modules, which does result in a tangible difference in performance.
The CAS latency is measured in clock cycles, and each clock cycle @ 533MHz is 25% longer than that of 667MHz. So CL4 @ 533MHz actually corresponds to the same absolute latency as CL5 @ 667MHz.
Posted: Thu Jan 04, 2007 7:19 am
by nxman
Austin you did a great job and this is an excellent post.
Posted: Thu Jan 04, 2007 8:11 am
by Jaduncan
I tested all XP 32bit & x64 edition, server 2003 Enterprise Edition, Vista Ultimate 32bit, the available RAM amounted to 3GB only, PAE was enabled though. Any idea how to fully utilize 4GB of RAM?
At the risk of sounding like a fanboi, if you really want to do this then Linux should be able to address all of it.
Posted: Sun Jan 07, 2007 6:13 am
by Austin_Goh
T2600 -> T7600
Posted: Mon Jan 08, 2007 8:08 am
by stefan_s1
After seeing this I orderd a T7600 today.... cant wait to have it replace my T2600
I will post my results here...

Posted: Mon Jan 08, 2007 12:51 pm
by harout
I ordered a refurbished T60p on ebay for $2500 or so. Why refurb? Why without any warranty? Because of the magnificent UXGA screen. IBM want $171/year for additional warranty!

And this seller sold me some Mack Computer international joke warranty thing.

May as well assume it has none.
Anyway, None of that wide stuff here

! I think I will wait maybe ~8 months or so for the price of the C2Ds to drop and when the extra performance will be most needed. I might even be tempted to throw a C2Q

if they release a mobile version with the same socket by then.
Performance with RAM Drive
Posted: Mon Jan 08, 2007 2:07 pm
by PaulS
How much of a performance increase does a RAM Drive provide?
Re: Performance with RAM Drive
Posted: Sun Jan 14, 2007 10:12 am
by Austin_Goh
PaulS wrote:How much of a performance increase does a RAM Drive provide?
Quantitatively, it is like about 1:40 speed enhancement over hard disk judging by the HD Speed Benchmark I posted earlier.
But I only installed frequently used programs on RAM Drive, how I wish the whole Hard Disk is as fast as the RAM Drive I created with utility.

Re: Performance with RAM Drive
Posted: Mon Jan 15, 2007 12:21 pm
by BillMorrow
Austin_Goh wrote:PaulS wrote:How much of a performance increase does a RAM Drive provide?
Quantitatively, it is like about 1:40 speed enhancement over hard disk judging by the HD Speed Benchmark I posted earlier.
But I only installed frequently used programs on RAM Drive, how I wish the whole Hard Disk is as fast as the RAM Drive I created with utility.

SO, then, unless the buss speed of the HDD is the limiting factor it might be interesting to consider a RAM based HDD with no moving parts..
just a shell filled with memory..
there must be something holding up development of a RAM based hard drive substitute..
Re: Performance with RAM Drive
Posted: Mon Jan 15, 2007 2:26 pm
by brentpresley
BillMorrow wrote:Austin_Goh wrote:
Quantitatively, it is like about 1:40 speed enhancement over hard disk judging by the HD Speed Benchmark I posted earlier.
But I only installed frequently used programs on RAM Drive, how I wish the whole Hard Disk is as fast as the RAM Drive I created with utility.

SO, then, unless the buss speed of the HDD is the limiting factor it might be interesting to consider a RAM based HDD with no moving parts..
just a shell filled with memory..
there must be something holding up development of a RAM based hard drive substitute..
They are out (upto 64GB), but EXTREMELY expensive. Like $1000 expensive.
They are also not that fast because they use NAND memory, not the memory we use for RAM (which is volatile and gets cleared when the computer turns off).
These drives are EXCELLENT for finding files (access time is < 0.1ms) but transfer speeds need work (currently in the 60-70MB/s range).
It is the same RAM used on those "high-speed" SD camera cards. Ever noticed how you can bring up a directory nearly instantly on those cards, but copying all the files somehow doesn't tax your hard drive (at least not mine)? Transfer speeds.
Those are the negatives (speed and cost). The positives are these things have no moving parts (good NAND can do about 70,000 read/write cycles) and power utilization is a FRACTION of what a standard hard drive is.
EDIT: here is a link for one:
http://www.pqi.com.tw/news_1.asp?ID=1444
Posted: Thu Jun 14, 2007 8:52 am
by Dbruyere
Austin_Goh, I have the 2007-EC1 which has the same stock specs as yours and after my upgrade to the T7200, the bios incorrectly identifies the processor I think. Could you tell my what your's shows? Mine is identified as:
Intel(R) Pentium(R) M CPU 000 @ 2.00GHz 2.00GHz
That is also what it says when I right click on my computer and choose properties. There is no mention of "Core" anywhere.
I running the latest BIOS and Windows Vista Business.
Thanks for any help!
Posted: Thu Jun 14, 2007 12:37 pm
by axiom
4GB RAM?
How can it be possible?
Posted: Thu Jun 14, 2007 3:26 pm
by RonS
axiom wrote:4GB RAM?
How can it be possible?
It's not possible. He will be able to see only 3GB of memory with the T60. If he had a T61, it would be a different story.