Re: Strange authorisation isuses:

"Dragan" ([email protected] via alfs-discuss Mailing List) <[email protected]>
Newsgroups gmane.linux.lfs.automated
Message-ID <[email protected]>
On 18/06/2022 00:01, Bruce Dubbs ([email protected] via alfs-discuss 
Mailing List) wrote:
> On 6/17/22 15:17, Dragan ([email protected] via alfs-discuss Mailing 
> List) wrote:
>>
>> On 17/06/2022 01:57, Bruce Dubbs ([email protected] via 
>> alfs-discuss Mailing List) wrote:
>>> On 6/16/22 16:04, Dragan ([email protected] via alfs-discuss Mailing 
>>> List) wrote:
>>>
>>>> Thanks for the scripts.
>>>>
>>>> About timing: SBU of 75 sec was based on -j1 and not -j16. I have 
>>>> set binutils to be built without optimisation.
>>>> On the same system in VM I got SBU between 20 aand 22 when building 
>>>> using my own scripts and that is inclusief unpacking source code 
>>>> and logging using -j16. But maintaining scripts between dev 
>>>> releases is just to much time consuming and mistake prone.
>>>>
>>>> On wsl using -j24 I got times between 18,9 and 19,5 sec using full 
>>>> scale virtualization available using the same script. But with wsl 
>>>> enabled is Virtual Box practically useless, it feels like "good old 
>>>> time" of 8080 and I have thrown it away.
>>>>
>>>> What have realy surprised me is surprisingly low time of Kernel 
>>>> build 93 seconds. I have already booted  in the system and it's 
>>>> looks ok. I will take deeper look at the issue in chroot.
>>>>
>>>> Total build time is not issue. Multitasking with some power to spare...
>>>
>>> That HW is interesting.  What is your output of
>>>
>>> $ lscpu | head -n18
>>>
>>> What I have is:
>>>
>>> Architecture:                    x86_64
>>> CPU op-mode(s):                  32-bit, 64-bit
>>> Address sizes:                   43 bits physical, 48 bits virtual
>>> Byte Order:                      Little Endian
>>> CPU(s):                          24
>>> On-line CPU(s) list:             0-23
>>> Vendor ID:                       AuthenticAMD
>>> Model name:                      AMD Ryzen 9 3900X 12-Core Processor
>>> CPU family:                      23
>>> Model:                           113
>>> Thread(s) per core:              2
>>> Core(s) per socket:              12
>>> Socket(s):                       1
>>> Stepping:                        0
>>> Frequency boost:                 enabled
>>> CPU max MHz:                     4672.0698
>>> CPU min MHz:                     2200.0000
>>> BogoMIPS:                        7585.83
>>>
>>>   -- Bruce
>>
>> I have today finished my fourth build with logging an all possible 
>> tests run once and al logging enabled duration just above three hours 
>> while I have used host quite intensively.
>>
>> But my fifth build while I have not used host at al finished in 124.1 
>> SBU of 19 sec as defined by binutils compiled with -j16. If I have 
>> didn't calculated wrongly that means build time of 39 min 30 sec 
>> build time. I will submit report to the list tomorrow.
>>
>> That was real unexpected (from my side) for VM. My virtual hardware 
>> spec is below.
>>
>> lscpu | head -n18
>> Architecture:                    x86_64
>> CPU op-mode(s):                  32-bit, 64-bit
>> Address sizes:                   46 bits physical, 48 bits virtual
>> Byte Order:                      Little Endian
>> CPU(s):                          16
>> On-line CPU(s) list:             0-15
>> Vendor ID:                       GenuineIntel
>> Model name:                      12th Gen Intel(R) Core(TM) i9-12900K
>> CPU family:                      6
>> Model:                           151
>> Thread(s) per core:              1
>> Core(s) per socket:              16
>> Socket(s):                       1
>> Stepping:                        2
>> BogoMIPS:                        6374.38
>> Flags:                           fpu vme de pse tsc msr pae mce cx8 
>> apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht 
>> syscall nx rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc 
>> cpuid tsc_known_freq pni pclmulqdq vmx ssse3 cx16 pcid sse4_1 sse4_2 
>> x2apic movbe popcnt aes xsave avx rdrand hypervisor lahf_lm abm 
>> 3dnowprefetch invpcid_single tpr_shadow vnmi flexpriority vpid 
>> fsgsbase avx2 invpcid rdseed clflushopt md_clear flush_l1d 
>> arch_capabilities
>> Virtualization:                  VT-x
>>
>> Hypervisor vendor:               KVM
>>
>> And on the host extract from CPU-Z report:
>>
>> Processors Information
>> ------------------------------------------------------------------------- 
>>
>> Socket 1            ID = 0
>>      Number of cores        16 (max 64)
>>      Number of threads    24 (max 24)
>>      Hybrid            yes, 2 coresets
>>      Core Set 0        P-Cores, 8 cores, 16 threads
>>      Core Set 1        E-Cores, 8 cores, 8 threads
>>      Manufacturer        GenuineIntel
>>      Name            Intel Core i9 12900K
>>      Codename        Alder Lake
>>      Specification        12th Gen Intel(R) Core(TM) i9-12900K
>>      Package         Socket 1700 LGA
>>      CPUID            6.7.2
>>      Extended CPUID        6.97
>>      Core Stepping        C0
>>      Technology        10 nm
>>      Tjmax            100.0 °C
>>      Core Speed        3187.2 MHz
>>      Stock frequency        3200 MHz
>>      Max frequency        5200 MHz
>>      Instructions sets    MMX, SSE, SSE2, SSE3, SSSE3, SSE4.1, 
>> SSE4.2, EM64T, VT-x, AES, AVX, AVX2, FMA3, SHA
>>      L1 Data cache        8 x 48 KB (12-way, 64-byte line) + 8 x 32 
>> KB (8-way, 64-byte line)
>>      L1 Instruction cache    8 x 32 KB (8-way, 64-byte line) + 8 x 64 
>> KB (8-way, 64-byte line)
>>      L2 cache        8 x 1.25 MB (10-way, 64-byte line) + 2 x 2 MB 
>> (16-way, 64-byte line)
>>      L3 cache        30 MB (12-way, 64-byte line)
>>      Max CPUID level        00000020h
>>      Max CPUID ext. level    80000008h
>>      FID/VID Control        yes
>>
>> Dragan
>>
>> PS: I have postponed build of new machine waiting for Alder Lake 
>> because of performance.
>
> Very interesting.  I'll have to consider getting one of those.
>
>   -- Bruce

My advice is to build your own machine if you can. I'm using Z690 UD 
DDR4 motherboard with 32 GB DDR4 memory

and old NVIDIA GeForce GTX 1050 Ti video card. I'm planing to switch to 
onboard Intel(R) UHD Graphics 770. More than enough firepower there for 
my purposes.

You can go for DDR5 memory for 2 to 5% increase in performance but I 
don't believe that price/performance ratio will be justified you better 
double the memory and invest difference in motherboard price in one or 
two fast hard drive.

Dragan

-- 
http://lists.linuxfromscratch.org/sympa/info/alfs-discuss
Unsubscribe: See the above information page
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.