| View previous topic :: View next topic |
| Author |
Message |
shrike
Apprentice
Joined: 20 Feb 2004
Posts: 187
Location: Closer to home
|
 Posted: Thu Feb 23, 2012 6:17 pm Post subject: AMD APU Pros/Cons? |
 |
|
I have been considering options for a move to a quad core system from a X2 dual core. Recent HW issues with the X2 plus a wish for UEFI/GPT have me looking now. I have considered Intel Sandy Bridge offerings but USE=sna seems too rough at this time. And now I see UEFI is available on AMD MBs. So what are the Pros/Cons with AMD APUs on Gentoo?
Pros:
- Integrated GPU/CPU
- UEFI now available
Cons:
- ??
Are open source graphics drivers for AMD/ATI stable? March=native ok on AMD APUs (GCC-4.5.3)?
Thanks,
shrike |
|
| Back to top |
|
 |
Ant P.
Watchman
Joined: 18 Apr 2009
Posts: 6920
|
| Posted: Thu Feb 23, 2012 6:59 pm Post subject: |
|
|
The open driver is usually stable, though pretty slow.
-march=native *should* work, but those chips don't have 3dnow instructions and I think they were released after GCC 4.5. You can find out what -march=native adds by doing:
| Code: | | gcc -v -march=native -x c /dev/null 2>&1 | grep -- '-march' | egrep -o -- '-+(m|param )\S+' |
|
|
| Back to top |
|
|
Jogie214
Apprentice
Joined: 15 Aug 2004
Posts: 159
Location: Cologne / Germany
|
| Posted: Fri Feb 24, 2012 1:18 am Post subject: |
|
|
Hi,
I've a E-350 running here, the use of -march=native gave mw a lot of problems with qt apps since gcc is using 3dnow support, which the processor doesn't provide  . Not sure about other AMD APUs though. The Situation is supposed to be resolved with the gcc-4.6.x.
I use these:
| Code: |
CFLAGS="-mtune=generic -mmmx -msse -msse2 -msse3 -mssse3 -msse4a -mabm -O2 -pipe -mcx16 -mpopcnt"
|
Greets,
Sebastian
_________________
Desktop: Ryzen7 1700x / x370 / Radeon RX 550 / Gentoo amd64 |
|
| Back to top |
|
|
shrike
Apprentice
Joined: 20 Feb 2004
Posts: 187
Location: Closer to home
|
| Posted: Fri Feb 24, 2012 4:20 am Post subject: |
|
|
Thnks for the feedback!
I should have been specific as to hardware which is the AMD FM1 socket with Llano APUs. I've searched the net and found a number of articles. This one,
http://www.phoronix.com/scan.php?page=article&item=amd_a83850_cpu&num=1
shows /proc/cpuinfo output for the AMD Fusion A8-3850 Llano.
| Code: |
processor : 3
vendor_id : AuthenticAMD
cpu family : 18
model : 1
model name : AMD A8-3850 APU with Radeon(tm) HD Graphics
stepping : 0
cpu MHz : 800.000
cache size : 1024 KB
physical id : 0
siblings : 4
core id : 1
cpu cores : 4
apicid : 3
initial apicid : 1
fpu : yes
fpu_exception : yes
cpuid level : 6
wp : yes
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 mmxext fxsr_opt pdpe1gb rdtscp lm 3dnowext 3dnow constant_tsc rep_good nopl nonstop_tsc extd_apicid aperfmperf pni monitor cx16 popcnt lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt npt lbrv svm_lock nrip_save pausefilter
bogomips : 5800.20
TLB size : 1536 4K pages
clflush size : 64
cache_alignment : 64
address sizes : 40 bits physical, 48 bits virtual
power management: ts ttp tm stc 100mhzsteps hwpstate
|
/proc/cpuinfo for my X2:
| Code: |
processor : 1
vendor_id : AuthenticAMD
cpu family : 15
model : 107
model name : AMD Athlon(tm) 64 X2 Dual Core Processor 4400+
stepping : 1
cpu MHz : 2310.414
cache size : 512 KB
physical id : 0
siblings : 2
core id : 1
cpu cores : 2
apicid : 1
initial apicid : 1
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
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 mmxext fxsr_opt rdtscp lm 3dnowext 3dnow rep_good nopl extd_apicid pni cx16 lahf_lm cmp_legacy svm extapic cr8_legacy 3dnowprefetch lbrv
bogomips : 4622.62
TLB size : 1024 4K pages
clflush size : 64
cache_alignment : 64
address sizes : 40 bits physical, 48 bits virtual
power management: ts fid vid ttp tm stc 100mhzsteps
|
The concept of combining CPU and GPU on one die thereby losing bulky/noisey graphics cards is very appealing to me.. if it works under Gentoo. Has anyone tried this hardware with gentoo?
Thanks,
shrike |
|
| Back to top |
|
|
Ion Silverbolt
Apprentice
Joined: 04 Nov 2004
Posts: 203
|
| Posted: Sat Feb 25, 2012 4:30 pm Post subject: |
|
|
| Jogie214 wrote: | Hi,
I've a E-350 running here, the use of -march=native gave mw a lot of problems with qt apps since gcc is using 3dnow support, which the processor doesn't provide . Not sure about other AMD APUs though. The Situation is supposed to be resolved with the gcc-4.6.x.
I use these:
| Code: |
CFLAGS="-mtune=generic -mmmx -msse -msse2 -msse3 -mssse3 -msse4a -mabm -O2 -pipe -mcx16 -mpopcnt"
|
Greets,
Sebastian |
March=native or march=btver1 works great with GCC 4.6.2 for the Zacate CPU's. I have a whole system successfully compiled using those settings. Even though GCC 4.6.2 is hard masked, it's very stable. I haven't found anything that won't compile yet.
As for the Llano APU's, I think they're too new for specific settings. Unlike Zacate, they still have 3dnow so there shouldn't be any problems with using amdfam10 and adding some of the missing optimizations. |
|
| Back to top |
|
|
Etal
Veteran
Joined: 15 Jul 2005
Posts: 1932
|
| Posted: Sat Feb 25, 2012 5:56 pm Post subject: |
|
|
| Ion Silverbolt wrote: | | March=native or march=btver1 works great with GCC 4.6.2 for the Zacate CPU's. I have a whole system successfully compiled using those settings. Even though GCC 4.6.2 is hard masked, it's very stable. I haven't found anything that won't compile yet. |
The only problems stopping gcc-4.6 from being unmasked is that it compiles broken grub and libmpeg2 (so if you watch out for these two and switch gcc if you ever need to compile them, everything should be fine.)
_________________
“And even in authoritarian countries, information networks are helping people discover new facts and making governments more accountable.”– Hillary Clinton, Jan. 21, 2010 |
|
| Back to top |
|
|
Ion Silverbolt
Apprentice
Joined: 04 Nov 2004
Posts: 203
|
| Posted: Sat Feb 25, 2012 8:56 pm Post subject: |
|
|
| That explains the hard mask. Somehow both those package work on my system, but maybe it's hit and miss depending on other circumstances that exist on a system. |
|
| Back to top |
|
|
.:chrome:.
Advocate
Joined: 19 Feb 2005
Posts: 4588
Location: Brescia, Italy
|
| Posted: Sun Feb 26, 2012 1:40 am Post subject: |
|
|
I use an AMD E-350 Processor (Asus E35M-1i) with these settings:
| Code: | CFLAGS="-march=native -mtune=amdfam10 -mmmx -msse -msse2 -mssse3 -msse4a -Os -funsafe-loop-optimizations -fsched-pressure -pipe"
LDFLAGS="-Wl,-O1 -Wl,--enable-new-dtags -Wl,--sort-common -Wl,-s" |
It works well as any Intel Atom I have tried! |
|
| Back to top |
|
|
|
Kernel & Hardware
