juandante

Note that if your sole purpose is to have a socketed laptop, any cheap i5 is faster and cooler and battery efficient than the best i7 socketed laptop back in the day. Now, one BIG advantage of socketed laptops is that they usually have a lot of internal and external ports and in consequence storage upgrade possibilities, that's why I still have the 2570p. You can upgrade a lot of things, one day make it powerful, another day make it lightweight and power efficient. It is a lot of fun to have control on your laptop. On other hands, if you can search very good, some rare HP/dell/lenovo etc professional/enterprise ultrabooks do offer the same number of internal and external ports, and they have bigger screen like 14-15 inch and almost identical phisical size as the 2570p because of the bezels and thickness of old laptops If you get a 2570p in 2023 it must be maximum 50 bucks, if you want to spend more than 100 bucks on any laptop still in 2023 you must get a recent at least a gen 8 (before gen 8 pretty much all Intel CPU where shitty only exception is the 45W laptop i7 but they where not socketed begining 4th Gen), or get an AMD (the recent AMD laptops after 2020 are exceptionnal, faster in all aspects and cheaper than Intel) The 2570p is still one of the best for 12 inch if you manage to get one for 50 bucks, as well as its lenovo counterpart (forgot the name but you can search this thread)

No need to upgrade heatsink. I am running a 3940xm since a year and the CPU will max out at 3.6 Ghz on 6 cores, 3.2 Ghz on 8 cores (obviously with HT, knowing that the max freq on 8 cores is 3.6 Ghz, but the BIOS is virtually limiting the multiplier on the XM editions to 3.2 Ghz). Means : I will run at 3.0 Ghz in stress test at ~95°C. So a 3840 QM will max out without problem at maybe 80-85°C or even 95°C because my XM is clamped. The solution ? Buy good thermal paste. Thermal Grizzly Kryonaut, I gained 10-15°C (maybe even more) only by swapping this thermal paste. Drastic change. I drilled a small hole (half the size of a penny). also on the back. I can even run fan off with good ThrottleStop settings.

+1 Very good. The original 3G is a pain in the *ss, I couldn't even watch a Yt video. It is hard to imagine people where using such slow 3G cards back in the day, lol. 3G+ is the bare minimum for modern internet. Also, for people looking for a replacement of the 2570p, the latest Dell Latitude 5521 are very good (they where out maybe last year or in 2020). I used it and it reminded me of the 2570p (multiple ports, etc). But the most impressive is that the Latitude is 15" is slightly bigger but it really feels as light and portable as the 2570p for all what it can do. Also, the CPU is 2x even 3x faster. Unfortunatly as we all know, newer CPUs are not swappable in newer laptops.

Please stop with this negligence. "Insider program" = "Beta". It is a marketing term. Running Beta software is risky towards final software. We saw this endless times. The real question, is the computer upgradable on TPM chip ?

I had a 3820QM and it was almost never throttling except when hot (100°). If no thermal throttle, it maxed out at arround 40W with 90W charger. I thing the 5W difference (b/c it is rated 45W) is for the GPU, even with TurboBoost the 3820QM was not exceeding 40W with good cooling and almost never throttled. See my previous posts. What about TPM 2.0 ? Is the 2570p compatible with Windows 11 natively, with mods ? Is the TPM 1.2 chip of the 2570p upgradable ?

So there is no thermal throttle with the 3940XM (and in consequence, all less power-hungry CPUs as the 3820QM, etc) ? What is the maximum CPU temps in ThrottleStop (screenshot) ? If there is no thermal throttling, this will be undoubtedly the best "simple" mod for the 2570p to unleash its full power. I did some basic calculations and the 3820QM if there was no throttling would bring the CPU to more than 120°C (at full power the CPU is throttling at about 25W on my laptop, while its normal power is 40W on the 2570p), I find it hard to believe that this cooler can bring down the laptop to 20°C (at least) to limit throttling. I tried a lot of impractical cooling methods (even the freezer, lol) and never managed to bring down the CPU to less than 100°C with no throttling in full load for more than a handful of minutes. How is this possible ? Can you post more information (room temperature, CPU temps, run time, screenshots) ?

Hello, I am wondering if someone already tried a vacuum cooler on the 2570p as seen below ? My goal is double : 1/ be able to use the laptop in silent mode (the default fan of the 2570p is not enough silent even at lower RPMs) so the vacuum cooler can "replace" the internal fan when the CPU is in low performance mode and 2/ reduce thermal throttling for the quad-cores in high performance mode (the default fan is not enough powerful, even with upgraded heatpipes). Also, what do you think about thermoelectric pads ? A thermoelectric pad is like a "mini freezer" (doesn't need any fan or water to cool down something, but needs far more power). My second choice would be to use this pad plugged on an external powersouce (USB seems not enough powerful). Also, there are pads specifically made for smartphones, obviously they will not be able to cool down a full fledge laptop CPU but they could help to reduce the heat in power-saving mode, with no external battery. Nevertheless, didn't find suitable ones for the laptop, and here is a standard thermoelectric pad : https://www.amazon.fr/KIMILAR-Dissipateur-Refroidisseur-thermoélectrique-refroidissement/dp/B00ITMJPCU/ref=sr_1_2?__mk_fr_FR=ÅMÅŽÕÑ&dchild=1&keywords=powerful+laptop+cooler+peltier&qid=1616708231&sr=8-2 And a vacuum cooler : https://www.amazon.com/LC05-Cooling-Auto-Temp-Detection-Compatible/dp/B00XKU47Y2/ref=as_li_ss_tl?keywords=opolar+laptop&qid=1581658306&sr=8-2&linkCode=sl1&tag=root0d5-20&linkId=70b64017e0966f958d12ef53d92908ed&language=en_US

Of course they will fit, it is the exact same laptop. The only difference between the two laptops is the light sensor removed on the 2570p on top of the screen.

Your benchmarks where interesting, but I observed the same lowering in battery life when I switched from dual core to quad core. I don't have my dual core anymore, to test again. I think that in your benchmarks the problem comes in full load, indeed, the dual core will be at full load of 2.4 Ghz (let's say), but the quad-core will also be at 2.4 Ghz in full load in Windows. When I say full load, I talk about medium load also (it is scarce to use a laptop in 100% idle, watching a film can be considered as idle but browsing Internet on the greddy Chrome, not very much). I would be interested to see the benchmark of system consumption with 2.4 Ghz (lets say) for the dual-core and quad-core. For sure it will be 2x more CPU consumption for the quad-core. Nevertheless, the quad-core should perform 2 times quicker than the dual-core, thus finishing the task 2 times earlier, but I doubt that Windows CPU governor is efficient to the point that it would turn off the corres as quickly as they stopped working. This would explain why I had this feeling also of the battery draining far more quickly on quad core than dual core, because in medium or full load, all the cores turns on (consuming 2x more power), and they don't turn off as quickly as they turn on. Edit : I was that you are on Linux Mint, maybe the governor is tweaked. But the reasoning on Windows (and thus, the consumption at 2.4 Ghz lets say, and full cores meaning 2c/4t for the dual core and consumption at 2.4 Ghz at 4c/8t should be still valid)

Edit

I would recommend you to go for the best. You have 16 Gb of RAM, so you need power. I have a 3820QM and I sometimes throttle but even with the throttle I have better benchmarks than 3610QM or dual core CPU (tested). If I could buy a 3940XM I would have done this. Contrary to what the first post says, in Windows any quad-core can be software limited to two-cores or one core, so if you need battery well this can easily be done in Windows. I will idle with this configuration dual-core mode on quad-core with 0.5W more than a real dual-core (tested between my original i5 35xxM and the i7 3820QM I have). This is negligeable towards all the speed advantages of the quad-core. All depends to you, do you want quad-core with negligable battery hit (I would not say 1 hour, more like 20-30 minutes with easily tweaking Windows, but this is nothing, if you need battery just buy a newer laptop), or do you want speed and a real work-station you can plug on an external monitor and have desktop like performance. Thus, this is less and less true with the latest Intel i7 10-gen processors, even i5 and AMD, caping at more than 4 Ghz.

The three big limiting factors of the 2570p according to me : 1/ Noisy with quad core CPU 2/ Screen colors are bad 3/ WebCam quality So, yes, I agree with you.

Yes it is really impressive, I was wishing for high quality mods as we can see on the Lenovo thinkpads Are you sure it is the 2570p ? Me I don't like PWM but I never had problems with screen flicker with the 2570p The low quality screen according to the very low grade color calibration was the main issue, all the time I used the 2570p

Would be pretty interesting to have a 2k screen, because with the 2570p + 3820qm and the low-res 768p screen I often feel like the power is not fully used.

Yes but no, there is no need to have a free cady for the sake of having a free cady. First you gain only few grams if it is a weight problem so this is pointless. The only useful thing with the mSATA is either RAID0+1, for high resilience, but a laptop with high resilience if pointless, or either having a 2TB hard drive and a high performance (800 to 1000 mb/s) RAID0 drive of SATA, this could be interesting. But you have still not told us about this question, is your drive in the WiFi slot or on the WWAN slot ? If you have 16Gb RAM, I recommend you to download the very old and known software "PrimoCache" it is a software to increase the cache of the SSDs (or HDDs) in using free RAM on the machine, so I have "RAM-like" speeds when loading some files (8 Gb/s for sequential read see benchs below). ----------------------------------------------------------------------- CrystalDiskMark 6.0.2 x64 (C) 2007-2018 hiyohiyo Crystal Dew World : https://crystalmark.info/ ----------------------------------------------------------------------- * MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s] * KB = 1000 bytes, KiB = 1024 bytes Sequential Read (Q= 32,T= 1) : 8450.388 MB/s Sequential Write (Q= 32,T= 1) : 6711.163 MB/s Random Read 4KiB (Q= 8,T= 8) : 1464.376 MB/s [ 357513.7 IOPS] Random Write 4KiB (Q= 8,T= 8) : 1273.470 MB/s [ 310905.8 IOPS] Random Read 4KiB (Q= 32,T= 1) : 412.207 MB/s [ 100636.5 IOPS] Random Write 4KiB (Q= 32,T= 1) : 471.074 MB/s [ 115008.3 IOPS] Random Read 4KiB (Q= 1,T= 1) : 444.689 MB/s [ 108566.7 IOPS] Random Write 4KiB (Q= 1,T= 1) : 413.750 MB/s [ 101013.2 IOPS] Test : 100 MiB [C: 60.9% (72.2/118.6 GiB)] (x1) [Interval=5 sec] Date : 2019/09/27 20:25:04 OS : Windows Server 2016 [10.0 Build 14393] (x64) PS : I have a mSATA drive on a standard SATA adapter in my 2570p so I could do the same screenshot as you without the BIOS mod ^^