**** BEGIN LOGGING AT Tue Jul 17 03:00:03 2018 Jul 17 03:05:57 Hurrian: Apparently lshw does not say which package type/form factor the CPU uses. :-/ Jul 17 03:37:04 IIRC you can't check the package type of the installed CPU from software, as socketed/soldered doesn't really make a difference to the software Jul 17 03:40:27 also, the microcode in the firmware probably isnt updated with the Ivy Bridge microcode, by then it was likely still in final qualification Jul 17 05:28:56 Hurrian: That makes sense but lots of things can be checked from software that usually should not matter to application software, such as the type of RAM or the chassis type of the computer. Jul 17 05:31:15 Only if the vendor declares those properties in e.g. ACPI tables Jul 17 05:31:21 or the DMI table Jul 17 05:35:17 Hurrian: Maybe a better example is the brand of CPU as reported by the CPUID instruction. How often does application software need to know on which brand of CPU it runs as long as the CPU does what the software needs? Jul 17 05:38:33 Maybe the government of the PRC wants to avoid VIA CPUs because VIA is in the ROC? Jul 17 05:39:47 often never, but iirc optimizing compilers sometimes enable faster code paths depending on available CPU features Jul 17 05:44:19 Well, the position of the CPC is that the ROC no longer legitimately exists, so I guess the CPC does not care if they use hardware from Taiwan because their position is that Taiwan is a province of the PRC. Jul 17 05:46:49 Hurrian: Features of the CPU running the compiler or of the CPU targetted by the compiler? Jul 17 05:47:32 the latter Jul 17 05:51:17 Hurrian: Then why do features of the CPU running the compiler matter? Use of cross-compilers is common, even within the same family of computer/instruction set architecture, such as using an x86-64 computer to build software for x86-32. Jul 17 05:52:07 the CPU running the compiler doesn't matter as long as it's set up to do cross compile properly Jul 17 05:52:35 for gcc at least, building cross-endian compilers more or less works too, for most architectures Jul 17 06:02:10 Hurrian: I guess non-GNU compilers have had cross-endian compilation working for decades too because, as far as I know, x86 (little endian) computers are used to run compilers targetting big-endian computers, such as embedded platforms using m68k or PowerPC. Jul 17 06:04:54 I guess a specific example is CodeWarrior or some other non-GNU compiler targetting an m68k Palm OS computer but running on an x86 computer. Jul 17 06:11:31 Or a Microsoft compiler running on an x86 computer but targetting a big-endian video game console, such as one using the MIPS (PlayStation, Nintendo 64) or SuperH/SH (Sega Saturn, Sega Dreamcast) architecture. I forgot the endianness of MIPS and SH but I remember that m68k is big-endian and PowerPC supports both little- and big-endian modes but Darwin/Mac OS X uses big-endian mode. Jul 17 06:16:54 If I recall correctly, Alpha also supports both little- and big-endian modes. Jul 17 06:18:45 If I recall correctly, 6800/6502 are little-endian, which seems unexpected because 68000 is big-endian. Jul 17 06:20:55 Does a Sun x86 computer use the IBM PC architecture or does Sun have their own computer architecture using the x86 ISA? Jul 17 06:24:19 There was the SGI Visual Workstation that, if I recall correctly, uses the x86 ISA but at least partly does not use the IBM PC architecture. Jul 17 22:00:08 Hurrian: Can you recommend a model of 8-GiB SODIMM that will work with this notebook computer with the current CPU? Jul 17 22:00:37 I mean with the CPU currently installed. Jul 17 22:10:56 Also, will this computer with the current CPU work with 2× 8-GiB SODIMMs? Jul 17 23:57:56 yep, would probably be better to buy matching SODIMMs you could use dual channel on all 16GB of memory Jul 17 23:58:50 as for a model of SODIMM, any 1333MHz 1.5V DDR3 SODIMM would work if you bought two, but if you're only buying one you need to match at least the primary timings Jul 18 00:31:39 Hurrian: Apparently both of the current 4-GiB SODIMMs are the same model but only one is easily accessible by the user. Assuming that both modules are removable, I guess accessing the other module requires disassembling the computer instead of only removing a door/cover from the bottom of the computer. Jul 18 00:32:11 it's probably accessible once the keyboard/palmrest are removed Jul 18 00:32:32 haven't really seen any laptops where you have to remove the motherboard to get to the DIMMs Jul 18 00:34:34 Some models of notebook computers, including an Asus model my colleague has, have the main memory soldered on the motherboard. I believe the Asus model is Asus X200MA-US01T. Jul 18 00:39:21 oh yeah, you have to check first if the preinstalled "DIMM" is actually soldered on the motherboard Jul 18 00:39:37 some laptops from that era have the main DIMM soldered, plus one expansion slot Jul 18 01:44:10 Hurrian: I meant main memory in the computer architecture sense, not in the computer hardware design sense, i.e., the Asus notebook computer uses DRAM ICs soldered on the motherboard instead of using a (SO)DIMM socket. Jul 18 01:44:33 yeah, newer laptops have all the RAM soldered in Jul 18 01:44:47 but some laptops from a few years ago have preinstalled soldered RAM, plus one expansion slot Jul 18 01:48:30 Hurrian: Yes, I knew what you meant but I realised that “main memory” was possibly ambiguous in this context. I avoid the common usage of saying “RAM” instead of “main memory” because a computer has RAM that is not part of the main memory, such as all of the RAM on all of the peripherals. Jul 18 01:51:58 “but some laptops from a few years ago have preinstalled soldered RAM, plus one expansion slot” Jul 18 01:51:58 I think this was done even in the 1990s, such as on some ThinkPad models. I think my Toshiba Satellite 2230CDS from 1999 or 2000 did that. Jul 18 01:53:28 I think my Dell Latitude X1 from 2005 is that way too. **** ENDING LOGGING AT Wed Jul 18 03:00:01 2018