1 |
J. Roeleveld wrote: |
2 |
> On December 8, 2018 6:23:04 PM UTC, Dale <rdalek1967@×××××.com> wrote: |
3 |
> |
4 |
> Alexander Puchmayr wrote: |
5 |
> |
6 |
> Am Donnerstag, 6. Dezember 2018, 10:27:31 CET schrieb Dale: |
7 |
> |
8 |
> Howdy, I mentioned in other threads that I'm doing some |
9 |
> upgrades to my system. My first question is about a CPU |
10 |
> upgrade. I currently have this for my CPU, from cpuinfo: |
11 |
> AMD Phenom(tm) II X4 955 Processor I've bought but not yet |
12 |
> installed a FX-8350 CPU. I have this in my make.conf file: |
13 |
> CFLAGS="-march=native -O2 -pipe" |
14 |
> |
15 |
> Compiling the whole system with -march=native might lead to |
16 |
> troubles, especially when doing a CPU change. This option |
17 |
> means that gcc is determining the type of CPU automatically |
18 |
> and adjusts the instruction set used to exactly this CPU. |
19 |
> Although, in your case, it is highly likely that your new CPU |
20 |
> understands all commands from the old, but I wouldn't bet on |
21 |
> it. Its possible that your existing software encounters |
22 |
> problems like "illegal instruction" or the like. Very bad if |
23 |
> your compiler crashes after CPU replacement, then you cannot |
24 |
> emerge anything. I highly recommend using CFLAGS="-O2 -pipe" |
25 |
> and nothing more, the performance difference is, if measurable |
26 |
> at all, negligible. |
27 |
> |
28 |
> USE_CPU="fpu vme de pse tsc msr pae mce cx8 apic sep mtrr |
29 |
> pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht |
30 |
> syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm 3dnowext |
31 |
> 3dnow constant_tsc rep_good nopl nonstop_tsc extd_apicid |
32 |
> pni monitor cx16 popcnt lahf_lm cmp_legacy svm extapic |
33 |
> cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs |
34 |
> skinit wdt nodeid_msr hw_pstate npt lbrv svm_lock nrip_save" |
35 |
> |
36 |
> As someone else in this thread already mentioned, USE_CPU is |
37 |
> not used. What you're looking for is CPU_FLAGS_X86=..., which |
38 |
> defines what cpu-specific options will be enabled for packages |
39 |
> supporting it and where it makes sense. See package |
40 |
> cpuid2cpuflags for details. Regards Alex |
41 |
> |
42 |
> |
43 |
> It seems the holiday shopping is slowing down delivery. My fan was |
44 |
> supposed to be here today but didn't arrive. Since I got time, I'll |
45 |
> change the CFLAGS for at least the @system stuff, that should get me |
46 |
> booted for sure. While the native setting makes things easier for |
47 |
> normal use, I can see the point of not using it when changing CPUs. |
48 |
> That is one reason for this thread. The CPUs are different and may |
49 |
> require some changes during the swap. |
50 |
> |
51 |
> Is there a easy way to see what if any changes will be made? I did a |
52 |
> emerge -UDNa @system but it's not showing any change. Does it require a |
53 |
> emerge -e @system to force the change? Or is it not changing anything? |
54 |
> |
55 |
> Thanks much. Better safe than sorry. ;-) |
56 |
> |
57 |
> Dale |
58 |
> |
59 |
> :-) :-) |
60 |
> |
61 |
> |
62 |
> A CFLAGS change requires a rebuild of all packages done with gcc. I am |
63 |
> not aware of a simple way of only doing those, so a "emerge --empty |
64 |
> @world" will be needed. |
65 |
> |
66 |
> -- |
67 |
> Joost |
68 |
> -- |
69 |
> Sent from my Android device with K-9 Mail. Please excuse my brevity. |
70 |
|
71 |
|
72 |
Based on the output, that's what I was thinking. Emerge picks up on |
73 |
other USE changes but it seems it only grabs the CFLAGS during the |
74 |
compile/configure phase for each package. Would this change the kernel |
75 |
image as well or would it remain the same? I may build a new kernel |
76 |
just to be sure. |
77 |
|
78 |
One good thing about this, I can compare the times with current CPU and |
79 |
new CPU later and get a rough idea of speed increases. ;-) |
80 |
|
81 |
Pardon me while I generate some heat. o_O |
82 |
|
83 |
Dale |
84 |
|
85 |
:-) :-) |