[gentoo-user] new computer : any advice ?

[gentoo-user] new computer : any advice ?

[Philip Webb]

I've built  4  machines starting in 2000 ,
all with parts bought from Canada Computers in downtown Toronto.
ANB1 + ANB3 had Intel CPUs, ANB2 + ANB4 have AMD CPUs
('ANB' = 'Aristotle's new body', Aristotle being my computer spirit (grin)).
ANB1 (2000) went to the recyclers long ago, ANB3 (2007) is a door-stop ;
ANB2 (2003) still wakes up, but can't talk to a present-day mouse ;
ANB4 (2012) has worked perfectly to date ; ANB2-4 all run Gentoo.

Since I currently have only  1  usable machine, ie no stand-by,
I want to build a new cutting-edge replacement with ANB4 becoming stand-by.

Below are what I put into ANB4 & what is offered today for ANB5 :
I've marked my 1st-take choices with a '*'.
Since my AMD machines have been more successful than the Intels,
I intend to use an AMD CPU again & anyway they seem to be cheaper & faster.
Gigabyte mobos are cheaper & it has been good in ANB4.
I've always used Kingston memory with no problems.
I want Nvidia graphix & the faster card seems a better choice.
SSD HDD DVD drives don't offer much choice, but these seem adequate.

I use the machine for typical desktop office + personal work :
no gaming, no movies, no music (I might try enabling sound in ANB5).
Efficient handling of Gentoo emerges is important.

I am a bit shocked at the big increase in price
for a smaller increase in overall performance,
but I can afford the parts as described ; prices in CAD, of course.

ANB4 :

  120823 CPU : AMD Bulldozer X4 FX-4170 4-Core 4,2 GHz 8 MB    :  129.99
           Item CPUA003075 ; Part FD4170FRGUBOX
  120823 Gigabyte GA-990FXA-UD3 990FX+SB950 1866/1600/1333/1066 : 144.99 M20
          4x PCI Express x16 Gigabit LAN
          6x SATA 6GB/s 2x eSATA 6GB/s 4x USB 3.0+ 14x USB 2.0 AM3+ ATX
           Item MBGA001655 ; Part GA-990FXA-UD3
         Direct replacement : 1 yr                             :   15.95
  120823 Memory : Kingston HyperX 4 GB DDR3 1600 MHz CL9       :   24.99
           Item RAMK003793 ; Part KHX1600C9D3/4G
  120823 Graphix : Asus GT520-1GD3-CSM Nv GT520 810 MHz 1 GB   :   54.99
           Item VCAS002280 ; Part GT520-1GD3-CSM
  120823 SSD : OCZ Vertex4 128 GB SATA3 R 560 MB/s W 430 MB/s  :  114.99
           Item SMOC001080 ; Part VTX4-25SAT3-128G
  120823 HDD : Seagate Barracuda SATA3 500 GB 16 MB 6 Gb/s     :   69.99
           Item HDSG002084 ; Part ST500DM002
         Ont recycle fee                                       :    0.05
  120823 Case : Thermal Master TC-102 ATX Tower 500 W          :   42.99
           
                                                      subtotal :  598.93 
                                                           tax :   77.86
                                                         total :  676.79

ANB5 :

  CPU : AMD :
  
    CPUA003145 FD8320FRHKBOX : Downtown 4 : $ 205
    X8 FX-8320 (125 W) 8-Core Socket AM3+  3.5 GHz   8 Mb Cache  32 nm

    CPUA003140 FD8350FRHKBOX : Downtown 6 : $ 239
    X8 FX-8350 (125 W) 8-Core Socket AM3+  4   GHz   8 Mb Cache  32 nm

    CPUA003445 FD837EWMHKBOX : Downtown 3 : $ 269
    X8 FX-8370E (95 W) 8-Core Socket AM3+  4.3 GHz   8 Mb Cache  32 nm

  * CPUA003454 FD8370FRHKBOX : Downtown 1 : $ 279
    X8 FX-8370 (125 W) 8-Core Socket AM3+  4.3 GHz  16 Mb Cache  32 nm

  Mobo : Asus :

    MBAS003960 M5A97 LE R2.0    : Downtown 5 : $  95
    ASUS M5A97 LE R2.0 Socket AM3+ AMD 970/SB950 Chipset
    Dual Channel DDR3 2133(O.C.) MHz, 2x PCI-Express x16
    GLAN, 6x SATA 6.0 Gb/s, 2x USB 3.0, 6x USB 2.0
    ATX

    MBAS003965 M5A97 R2.0       : Downtown 2 : $ 130
    ASUS M5A97 R2.0 Socket AM3+ AMD 970/SB950 Chipset
    Dual Channel DDR3 2133(O.C.) MHz, 2x PCI-Express x16
    GLAN, 6x SATA 6.0 Gb/s, 2x USB 3.0, 6x USB 2.0
    ATX

    MBAS004035 M5A99X EVO R2.0  : Downtown 2 : $ 145
    ASUS M5A99X EVO R2.0 Socket AM3+ AMD 990X/SB950 Chipset
    Dual Channel DDR3 2133(O.C.) MHz, 3x PCI-Express x16
    GLAN, 6x SATA 6.0 Gb/s, 2x eSATA 6.0Gb/s, 2x USB 3.0, 8x USB 2.0
    ATX

    MBAS003885 M5A99FX PRO R2.0 : Downtown 3 : $ 160
    ASUS M5A99FX PRO R2.0 Socket AM3+ AMD 990FX/SB950 Chipset
    Dual Channel DDR3 2133(O.C.) MHz, 4x PCI-Express x16
    GLAN, 7x SATA 6.0 Gb/s, 1x eSATA, 2x USB 3.0, 8x USB 2.0
    ATX

    MBAS003995 SABERTOOTH 990FX R2.0 : Downtown 2 : $ 240
    ASUS TUF SABERTOOTH 990FX R2.0 Socket AM3+ AMD990FX/SB950 Chipset
    Dual Channel DDR3 1866 MHz, 4x PCI-Express x16
    GLAN, 8x SATA 6.0 Gb/s, 2x eSATA 6.0Gb/s, 4x USB 3.0, 8x USB 2.0
    ATX

    Gigabyte :

    MBGA002151 GA-970A-D3P : Downtown 4 : $ 105
    GIGABYTE GA-970A-D3P (rev. 1.0) Socket AM3+ AMD 970+SB950 Chipset
    Dual Channel DDR3 2000(O.C.) MHz, 2x PCI-Express x16
    GLAN, 6x SATA 6.0 Gb/s, 4x USB 3.0, 14x USB 2.0
    ATX

  * MBGA002794 GA-970A-UD3P : Downtown 6 : $ 120
    GIGABYTE GA-970A-UD3P (rev. 1.0) Socket AM3+ AMD 970+SB950 Chipset
    Dual Channel DDR3 2000(O.C.) MHz, 2x PCI-Express x16
    GLAN, 6x SATA 6.0 Gb/s, 4x USB 3.0, 14x USB 2.0
    ATX

    MBGA001655 GA-990FXA-UD3 : Downtown 3 : $ 150
    GIGABYTE GA-990FXA-UD3 Socket AM3+ AMD 990FX+SB950 Chipset
    Dual Channel DDR3 2000(O.C.) MHz, 4x PCI-Express x16
    GLAN, 6x SATA 6.0 Gb/s, 2x eSATA 6.0Gb/s, 2x USB 3.0, 8x USB 2.0
    ATX

  Memory :

    RAMK007183 HX318C10FRK2/8 : Downtown 6 : $  67
    Kingston HyperX Fury Red 8 GB (2 x 4 GB) DDR3 1866MHz CL10

    RAMK007147 HX318C10FK2/8 : Downtown 1 : $  67
    Kingston HyperX Fury Blue 8 GB (2 x 4 GB) DDR3 1866MHz CL10 DIMMs

  * RAMK007272 HX318C10FBK2/16 : Downtown 2 : $ 119
    Kingston HyperX Fury Black 16 GB (2 x 8 GB) DDR3 1866MHz CL10 DIMMs

    RAMK007281 HX318C10FRK2/16 : Downtown 6 : $ 120
    Kingston HyperX Fury Red 16 GB (2 x 8 GB) DDR3 1866MHz CL10 DIMMs

    RAMK007263 HX318C10FK2/16 : Downtown 1 : $ 120
    Kingston HyperX Fury Blue 16 GB (2 x 8 GB) DDR3 1866MHz CL10 DIMMs

  Graphix :

    VCAS002435 GT610-2GD3-CSM : Downtown 2 : $  65
    ASUS GeForce GT 610 2 GB DDR3
    810 MHz Clock, 1200 MHz Memory
    PCI Express 2.0, D-Sub, DVI, HDMI

    VCAS003034 GT730-SL-1GD3-BRK : Downtown 3 : $  85
    ASUS GeForce GT 730 1 GB GDDR3
    902 MHz Clock, 900 MHz Memory
    PCI Express 2.0, D-Sub/DVI/HDMI

    VCAS002883 GTX750-PHOC-1GD5 : Downtown 2 : $ 150
    ASUS GeForce GTX 750 1 GB GDDR5
    1059 MHz Clock, 5010 MHz Memory
    PCI-Express 3.0, DVI, HDMI

  * VCAS002892 GTX750TI-OC-2GD5 : Downtown 2 : $ 170
    ASUS GeForce GTX 750 Ti 2 GB GDDR5
    1072 MHz Clock, 5400 MHz Memory
    PCI-Express 3.0, DVI, HDMI

  SSD :

  * SMKT001725 SV300S37A/240GCR : Downtown 10 : $ 110
    Kingston SSDNow V300 240 GB 7 mm SATA 6 Gb/s  Read: 450 Write: 450 MB/s 

  HDD :

    HDSG002093 ST1000DM003 : Downtown 10 : $  66
    Seagate Desktop HDD 1 TB 3.5in SATA3 64 MB Cache OEM

  * HDSG003374 ST1000DM003 : Downtown 10 : $  58
    Seagate Desktop HDD 1 TB 3.5in SATA3 64 MB Cache OEM

    HDSG002084 ST500DM002 : Downtown 5 : $ 60
    Seagate Desktop HDD 500 GB 3.5in SATA3 16 MB Cache OEM

  DVD :

  * DVSS004328 SH-224FB/BSBE : Downtown 4 : $ 22
    Samsung (SH-224FB/BSBE) Internal 24x DVD Writer, OEM
    Black, SATA, 1.5 MB Buffer

-- 
========================,,============================================
SUPPORT     ___________//___,   Philip Webb
ELECTRIC   /] [] [] [] [] []|   Cities Centre, University of Toronto
TRANSIT    `-O----------O---'   purslowatchassdotutorontodotca

Re: [gentoo-user] new computer : any advice ?

[Dale]

Philip Webb wrote:
>   * MBGA002794 GA-970A-UD3P : Downtown 6 : $ 120
>     GIGABYTE GA-970A-UD3P (rev. 1.0) Socket AM3+ AMD 970+SB950 Chipset
>     Dual Channel DDR3 2000(O.C.) MHz, 2x PCI-Express x16
>     GLAN, 6x SATA 6.0 Gb/s, 4x USB 3.0, 14x USB 2.0
>     ATX


I have this mobo in my current rig.  So far, no problems.  I've had it
for a few years now.  It has a 4 core 3.2 GHz CPU in it and 16GBs of
ram.  I put it in a Cooler Master HAF-932 case, very large full tower
with HUGE fans.  It can hold 32GBs of ram tho.  I plan to upgrade one of
these days.  The mobo I upgraded from maxed out with 16GBs.  I just
moved them over since the 8GB sticks were pricey at the time, and not
much better now. 

Just in case you end up with this one, here is a list of the output of
lspci -k, which shows what kernel drivers it needs to work. 

root@fireball / # lspci -k
00:00.0 Host bridge: Advanced Micro Devices, Inc. [AMD/ATI] RD890 PCI to
PCI bridge (external gfx0 port B) (rev 02)
        Subsystem: Advanced Micro Devices, Inc. [AMD/ATI] RD890 PCI to
PCI bridge (external gfx0 port B)
00:00.2 IOMMU: Advanced Micro Devices, Inc. [AMD/ATI] RD990 I/O Memory
Management Unit (IOMMU)
        Subsystem: Advanced Micro Devices, Inc. [AMD/ATI] RD990 I/O
Memory Management Unit (IOMMU)
00:02.0 PCI bridge: Advanced Micro Devices, Inc. [AMD/ATI] RD890 PCI to
PCI bridge (PCI express gpp port B)
        Kernel driver in use: pcieport
00:04.0 PCI bridge: Advanced Micro Devices, Inc. [AMD/ATI] RD890 PCI to
PCI bridge (PCI express gpp port D)
        Kernel driver in use: pcieport
00:09.0 PCI bridge: Advanced Micro Devices, Inc. [AMD/ATI] RD890 PCI to
PCI bridge (PCI express gpp port H)
        Kernel driver in use: pcieport
00:11.0 SATA controller: Advanced Micro Devices, Inc. [AMD/ATI]
SB7x0/SB8x0/SB9x0 SATA Controller [AHCI mode] (rev 40)
        Subsystem: Gigabyte Technology Co., Ltd SB7x0/SB8x0/SB9x0 SATA
Controller [AHCI mode]
        Kernel driver in use: ahci
00:12.0 USB controller: Advanced Micro Devices, Inc. [AMD/ATI]
SB7x0/SB8x0/SB9x0 USB OHCI0 Controller
        Subsystem: Gigabyte Technology Co., Ltd SB7x0/SB8x0/SB9x0 USB
OHCI0 Controller
        Kernel driver in use: ohci-pci
00:12.2 USB controller: Advanced Micro Devices, Inc. [AMD/ATI]
SB7x0/SB8x0/SB9x0 USB EHCI Controller
        Subsystem: Gigabyte Technology Co., Ltd SB7x0/SB8x0/SB9x0 USB
EHCI Controller
        Kernel driver in use: ehci-pci
00:13.0 USB controller: Advanced Micro Devices, Inc. [AMD/ATI]
SB7x0/SB8x0/SB9x0 USB OHCI0 Controller
        Subsystem: Gigabyte Technology Co., Ltd SB7x0/SB8x0/SB9x0 USB
OHCI0 Controller
        Kernel driver in use: ohci-pci
00:13.2 USB controller: Advanced Micro Devices, Inc. [AMD/ATI]
SB7x0/SB8x0/SB9x0 USB EHCI Controller
        Subsystem: Gigabyte Technology Co., Ltd SB7x0/SB8x0/SB9x0 USB
EHCI Controller
        Kernel driver in use: ehci-pci
00:14.0 SMBus: Advanced Micro Devices, Inc. [AMD/ATI] SBx00 SMBus
Controller (rev 42)
        Subsystem: Advanced Micro Devices, Inc. [AMD/ATI] SBx00 SMBus
Controller
        Kernel driver in use: piix4_smbus
00:14.2 Audio device: Advanced Micro Devices, Inc. [AMD/ATI] SBx00
Azalia (Intel HDA) (rev 40)
        Subsystem: Gigabyte Technology Co., Ltd SBx00 Azalia (Intel HDA)
        Kernel driver in use: snd_hda_intel
00:14.3 ISA bridge: Advanced Micro Devices, Inc. [AMD/ATI]
SB7x0/SB8x0/SB9x0 LPC host controller (rev 40)
        Subsystem: Advanced Micro Devices, Inc. [AMD/ATI]
SB7x0/SB8x0/SB9x0 LPC host controller
00:14.4 PCI bridge: Advanced Micro Devices, Inc. [AMD/ATI] SBx00 PCI to
PCI Bridge (rev 40)
00:14.5 USB controller: Advanced Micro Devices, Inc. [AMD/ATI]
SB7x0/SB8x0/SB9x0 USB OHCI2 Controller
        Subsystem: Gigabyte Technology Co., Ltd SB7x0/SB8x0/SB9x0 USB
OHCI2 Controller
        Kernel driver in use: ohci-pci
00:16.0 USB controller: Advanced Micro Devices, Inc. [AMD/ATI]
SB7x0/SB8x0/SB9x0 USB OHCI0 Controller
        Subsystem: Gigabyte Technology Co., Ltd SB7x0/SB8x0/SB9x0 USB
OHCI0 Controller
        Kernel driver in use: ohci-pci
00:16.2 USB controller: Advanced Micro Devices, Inc. [AMD/ATI]
SB7x0/SB8x0/SB9x0 USB EHCI Controller
        Subsystem: Gigabyte Technology Co., Ltd SB7x0/SB8x0/SB9x0 USB
EHCI Controller
        Kernel driver in use: ehci-pci
00:18.0 Host bridge: Advanced Micro Devices, Inc. [AMD] Family 10h
Processor HyperTransport Configuration
00:18.1 Host bridge: Advanced Micro Devices, Inc. [AMD] Family 10h
Processor Address Map
00:18.2 Host bridge: Advanced Micro Devices, Inc. [AMD] Family 10h
Processor DRAM Controller
00:18.3 Host bridge: Advanced Micro Devices, Inc. [AMD] Family 10h
Processor Miscellaneous Control
        Kernel driver in use: k10temp
00:18.4 Host bridge: Advanced Micro Devices, Inc. [AMD] Family 10h
Processor Link Control
01:00.0 VGA compatible controller: NVIDIA Corporation GT216 [GeForce GT
220] (rev a2)
        Subsystem: NVIDIA Corporation GT216 [GeForce GT 220]
        Kernel driver in use: nvidia
        Kernel modules: nvidia
01:00.1 Audio device: NVIDIA Corporation GT216 HDMI Audio Controller
(rev a1)
        Subsystem: NVIDIA Corporation GT216 HDMI Audio Controller
        Kernel driver in use: snd_hda_intel
02:00.0 USB controller: VIA Technologies, Inc. Device 3483 (rev 01)
        Subsystem: Gigabyte Technology Co., Ltd Device 5007
        Kernel driver in use: xhci_hcd
03:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd.
RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller (rev 06)
        Subsystem: Gigabyte Technology Co., Ltd Motherboard
        Kernel driver in use: r8169
04:06.0 Ethernet controller: Davicom Semiconductor, Inc. Ethernet 100/10
MBit (rev 31)
        Subsystem: ARCHTEK TELECOM Corp Ethernet 100/10 MBit
        Kernel driver in use: dmfe
root@fireball / #


The short version:


root@fireball / # lspci -k | grep Kernel
        Kernel driver in use: pcieport
        Kernel driver in use: pcieport
        Kernel driver in use: pcieport
        Kernel driver in use: ahci
        Kernel driver in use: ohci-pci
        Kernel driver in use: ehci-pci
        Kernel driver in use: ohci-pci
        Kernel driver in use: ehci-pci
        Kernel driver in use: piix4_smbus
        Kernel driver in use: snd_hda_intel
        Kernel driver in use: ohci-pci
        Kernel driver in use: ohci-pci
        Kernel driver in use: ehci-pci
        Kernel driver in use: k10temp
        Kernel driver in use: nvidia
        Kernel modules: nvidia
        Kernel driver in use: snd_hda_intel
        Kernel driver in use: xhci_hcd
        Kernel driver in use: r8169
        Kernel driver in use: dmfe
root@fireball / #


On the CPU, I always get the 2nd or 3rd one down in price.  That is
usually the best bang for the buck.  You may not be able to tell a
difference between the 4.0GHz and the 4.3GHz one.  Then again, with 8
cores, it may be worth the extra few bucks to you.  That extra cache
could make a difference I guess.  All depends on what you can afford and
how long you plan to run the new rig. 

Hope this little bit of info helps you pick.  Maybe someone else will
post about what they have that you listed.  ;-)

Dale

:-)  :-) 

[gentoo-user] Re: new computer : any advice ?

[James]

Philip Webb <purslow <at> ca.inter.net> writes:

> Since my AMD machines have been more successful than the Intels,
> I intend to use an AMD CPU again & anyway they seem to be 
> cheaper & faster. Gigabyte mobos are cheaper & it has been good in ANB4.

I built some FX-8350 system on Gigabyte mobo (990-FXA) sata 3 and they
are just wonderful. I did put water coolers and a radiator on them
to keep the CPUs super cool. I have not over clocked them (yet) but 
it is all set up. Folks often overclock these rigs at 5 or 6 GHz, from
what I read. 32 gig of ram is sweet on these systems, at a
very attractive price point. I opted for max ram, vs a SSD. ymmv.


> I am a bit shocked at the big increase in price
> for a smaller increase in overall performance,
> but I can afford the parts as described ; prices in CAD, of course.


Check the newegg pricing as a reference.

Video:: AMD/ATI has a new smoking video card out that is tops in
memory bandwidth. Consider the fact that GCC-5.2+ allows one to 
compile and use the processor and memory resources of the hi end
video card (nvidia, intel, amd and other in this open standard);
I give some deep thought into putting those discretionary funds
into an advanced video card [1]. For sure this is still in the
developmental stages, but tons of ddr5 memory added to the system
resources is going to be a game changer once things stabilize. 


So I'm not really pushing any sort of hardware. But things are changing
with gcc-5.x so unless you are running big database/web sorts of codes, a
smaller amount of faster memory might just be the best performance/cost.
Max out the mobo ram first, with the fastest memory it can handle. ymmv.

hth,
James


[1] http://www.newegg.com/Product/Product.aspx?Item=N82E16814125797. 

Re: [gentoo-user] Re: new computer : any advice ?

[Stefan G. Weichinger]

Am 2015-09-08 um 17:32 schrieb James:

> So I'm not really pushing any sort of hardware. But things are changing
> with gcc-5.x 

could you detail this a bit?
Does it need less RAM or what?

I also consider getting a new system later this year, maybe a
core-i7-6xxx or so ...

Re: [gentoo-user] Re: new computer : any advice ?

[Francisco Ares]

[-- Attachment #1: Type: text/plain, Size: 687 bytes --]

2015-09-08 12:51 GMT-03:00 Stefan G. Weichinger <lists@xunil.at>:

> Am 2015-09-08 um 17:32 schrieb James:
>
> > So I'm not really pushing any sort of hardware. But things are changing
> > with gcc-5.x
>
> could you detail this a bit?
> Does it need less RAM or what?
>
> I also consider getting a new system later this year, maybe a
> core-i7-6xxx or so ...
>
>
>

Just my 2 cents:

IMHO, high end video cards, OpenCL capable (CUDA for nVidia), will be more
and more used for intensive and massive data manipulation, like audio and
video (image) recognition, which will be (I guess) the base of the future
user interfaces.

So why not get ready right now?  ;-)

Best regards,
Francisco

[-- Attachment #2: Type: text/html, Size: 1287 bytes --]

[gentoo-user] Re: new computer : any advice ?

[James]

Stefan G. Weichinger <lists <at> xunil.at> writes:


> > So I'm not really pushing any sort of hardware. But things are changing
> > with gcc-5.x 

I have posted several links on the subject previously [1]; here's one [2].
Basically several projects and the codes got pulled into 
GCC-5, where the GPU (processor and memory) resources
are utilized along with the systems CPU and ram for gcc-5.x
compiling and executing of codes.  DDR-5 for many algorithms
will make certain codes very fast to run.

> could you detail this a bit?
> Does it need less RAM or what?
As always more processors and more ram is better. Intel is already
moving higher end systems to DDR-4 on the mobo. Using the GPU
(SIMD) and the DDR-5 ram only adds to the capability of the machines.



> I also consider getting a new system later this year, maybe a
> core-i7-6xxx or so ...

If you can afford it, get a mobo that supports DDR-4.
Right now the AMD-HBM Fury-X is the video card with the 
highest bandwidth for a memory buss on a video card, if
you can find one for sale:: limited production right now.

RDMA Remote Dynamic Memory Access is the principal finally available
in gcc....

Game_changer imho.


hth,
James



hth,
James

[1] April 21 gentoo-user thread on GCC-5


[2] https://gcc.gnu.org/wiki/OpenACC

Re: [gentoo-user] Re: new computer : any advice ?

[Stefan G. Weichinger]

Am 2015-09-09 um 04:20 schrieb James:

> I have posted several links on the subject previously [1]; here's one [2].

[..]

> If you can afford it, get a mobo that supports DDR-4.
> Right now the AMD-HBM Fury-X is the video card with the
> highest bandwidth for a memory buss on a video card, if
> you can find one for sale:: limited production right now.
>
> RDMA Remote Dynamic Memory Access is the principal finally available
> in gcc....

Thanks for the pointers, I will read through that thread soon.
So this means chosing CPU *and* GPU accordingly :-)

I didn't plan to buy a separate video card at all as my usage is quite 
office/terminal-style without gaming or video stuff. The integrated 
graphics of modern core-i7xxx should be enough to run my 2 
24-inch-monitors. But if the GPU helps speeding up things ... I have to 
consider this as well.

Digging up that thread now somewhere ...

Stefan

Re: [gentoo-user] Re: new computer : any advice ?

[Stefan G. Weichinger]

On 09.09.2015 20:01, Stefan G. Weichinger wrote:
> Digging up that thread now somewhere ...

Ah, I even participated then ;-)

AFAI understand gcc-5 should compile faster?
And generate faster code in some cases?

looking at genlop -t I can't really spot speedups in the last months.

With gcc-5.2.0 as I have here right now:

5.2.0(5.2)^s(20:52:29 18.07.2015)(cxx fortran multilib nls openmp
sanitize -altivec -awt -cilk -debug -doc -fixed-point -gcj -go -graphite
-hardened -libssp -multislot -nopie -nossp -nptl -objc -objc++ -objc-gc
-regression-test -vanilla)

... does it use this new stuff anyway, do we need a specific USE-flag
enabled (I can't spot it, looking for something like "acc" or "rdma"
;-)), do we need specific CFLAGS .. ?

just curious.

btw: might be a projection but for some times back I noticed that my
gentoo installation got somehow "snappier" in a way. Might relate to my
rebuilding with gcc-5.x (and overall improvement of hundreds of
packages, sure).

Re: [gentoo-user] Re: new computer : any advice ?

[Mick]

[-- Attachment #1: Type: Text/Plain, Size: 1937 bytes --]

On Wednesday 09 Sep 2015 19:01:24 Stefan G. Weichinger wrote:
> Am 2015-09-09 um 04:20 schrieb James:
> > I have posted several links on the subject previously [1]; here's one
> > [2].
> 
> [..]
> 
> > If you can afford it, get a mobo that supports DDR-4.
> > Right now the AMD-HBM Fury-X is the video card with the
> > highest bandwidth for a memory buss on a video card, if
> > you can find one for sale:: limited production right now.
> > 
> > RDMA Remote Dynamic Memory Access is the principal finally available
> > in gcc....
> 
> Thanks for the pointers, I will read through that thread soon.
> So this means chosing CPU *and* GPU accordingly :-)
> 
> I didn't plan to buy a separate video card at all as my usage is quite
> office/terminal-style without gaming or video stuff. The integrated
> graphics of modern core-i7xxx should be enough to run my 2
> 24-inch-monitors. But if the GPU helps speeding up things ... I have to
> consider this as well.
> 
> Digging up that thread now somewhere ...
> 
> Stefan

I built last Christmas a Kaveri APU based PC, with two 23" monitors and no 
external GPU.  This is a PC used as a workstation for coding, image processing 
and the odd video transcoding.  No gaming.  Using stable radeon driver.  The 
performance of this machine has really impressed me when compiling packages, 
but I don't have the latest generation i7 to compare it with.  Unlike noisy 
discrete GPUs this thing is really quiet and doesn't consume much power 
either.  The Asus MoBo has a port for an external GPU, but I don't think I 
will ever bother getting one - certainly not new.  ;-)

I have been thinking that the better compiling performance compared to my 
other older PCs can't just be CPU specific, it must be all these additional 
HSA-enabled compute cores that are doing some of the heavy lifting.  I don't 
know how to check this though.

-- 
Regards,
Mick

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

Re: [gentoo-user] Re: new computer : any advice ?

[Jeremi Piotrowski]

On Wed, 9 Sep 2015, Stefan G. Weichinger wrote:

> On 09.09.2015 20:01, Stefan G. Weichinger wrote:
> > Digging up that thread now somewhere ...
>
> Ah, I even participated then ;-)
>
> AFAI understand gcc-5 should compile faster?
> And generate faster code in some cases?

No, and yes. Compilation is not affected in any way and runtime
performance can only be improved _if_ this stuff is explicitly used within
the code.

Meaning you would feel a difference in no less then 5 years when gcc-6 is
widely used and accelerator support is not restricted to intel MIC and
nvidia gpus. James is getting a bit ahead of himself calling this a
"game changer" - yeaaaaah... not really right now.

Right now this functionality is a toy for the HPC community and will stay
that way. To use it you have to build a separate offloading compiler, need
custom code used by few, and expensive hardware. The tree ebuild doesn't
even provide a way for enabling the accelerator support.


>
> ... does it use this new stuff anyway, do we need a specific USE-flag
> enabled (I can't spot it, looking for something like "acc" or "rdma"
> ;-)), do we need specific CFLAGS .. ?
>
> just curious.

I can't speak for RDMA (can't find any mention of it in gcc) because
that's an even more exotic thing than plain old accelerator support
(unless you run infiniband at home...), but the flags are:

    -fopenmp
    -foffload
    -fopenacc

However enabling them is as useful as having CFLAGS=-fopenmp currently. It
changes __nothing__ unless an application has openmp annotations, and the
ones that do should already provide a means of doing so in the build
system.

tldr: don't buy a dedicated gpu just because you read something on a
mailing list ;)

Re: [gentoo-user] Re: new computer : any advice ?

[Alan McKinnon]

On 09/09/2015 20:01, Stefan G. Weichinger wrote:
> Am 2015-09-09 um 04:20 schrieb James:
> 
>> I have posted several links on the subject previously [1]; here's one
>> [2].
> 
> [..]
> 
>> If you can afford it, get a mobo that supports DDR-4.
>> Right now the AMD-HBM Fury-X is the video card with the
>> highest bandwidth for a memory buss on a video card, if
>> you can find one for sale:: limited production right now.
>>
>> RDMA Remote Dynamic Memory Access is the principal finally available
>> in gcc....
> 
> Thanks for the pointers, I will read through that thread soon.
> So this means chosing CPU *and* GPU accordingly :-)
> 
> I didn't plan to buy a separate video card at all as my usage is quite
> office/terminal-style without gaming or video stuff. The integrated
> graphics of modern core-i7xxx should be enough to run my 2
> 24-inch-monitors. But if the GPU helps speeding up things ... I have to
> consider this as well.

I've done exactly that on my i7 laptop for ages now. It usually runs
it's own 1920 display and an external monitor of same resolution, and
has this hardware:

00:02.0 VGA compatible controller: Intel Corporation 4th Gen Core
Processor Integrated Graphics Controller (rev 06)
01:00.0 VGA compatible controller: Advanced Micro Devices, Inc.
[AMD/ATI] Venus XT [Radeon HD 8870M / R9 M270X/M370X]

I use the intel video driver and it's been more than a year since I
built a kernel with radeon :-)

The intel gpu manages full HD video and funky plasma5 effects just fine
at a fraction of the battery usage of the radeon. These days I ask
myself: if I'm not gaming at insane frame rates, or using cuda, then why
do I need the radeon power at all....?



-- 
Alan McKinnon
alan.mckinnon@gmail.com

Re: [gentoo-user] Re: new computer : any advice ?

[Stefan G. Weichinger]

On 09.09.2015 21:35, Jeremi Piotrowski wrote:

> tldr: don't buy a dedicated gpu just because you read something on a
> mailing list ;)

yep :-)

[gentoo-user] Re: new computer : any advice ?

[james]

Jeremi Piotrowski <jeremi.piotrowski <at> gmail.com> writes:

> No, and yes. Compilation is not affected in any way and runtime
> performance can only be improved _if_ this stuff is explicitly used within
> the code.

Yes this is all new and a work in progress. I do not think it will be
gcc-6 that makes the difference in a few years. But folks should be aware
and look for codes that are accelerated via usage of GPU resources.
Remember this all started about hardware purchase and future benefits.
It's definitely not commodity usage atm.


> Meaning you would feel a difference in no less then 5 years when gcc-6 is
> widely used and accelerator support is not restricted to intel MIC and
> nvidia gpus. James is getting a bit ahead of himself calling this a
> "game changer" - yeaaaaah... not really right now.

It's not as restricted as you indicate amd, intel, nividia and others like
arm (Mali and such) are working to support there hardware under the openacc
code extension now in gcc-5. Granted the more powerful your GPU resources
are the more they can contribute. This stuff use to only work with
vendor supplied compilers and sdks, now it's finally available in gcc,
albeit in it's infancy. Naturally it's going to take a while to 
become mainstream useful; but that more like a year or 2, at most.


> Right now this functionality is a toy for the HPC community and will stay
> that way. To use it you have to build a separate offloading compiler, need
> custom code used by few, and expensive hardware. The tree ebuild doesn't
> even provide a way for enabling the accelerator support.

I never said it was for the entire portage tree to benefit at this time,
but ultimately that is the goal. And yes this is the focus of the cluster
folks, including HPC, but it will soon be available to everybody, via
clusters and distributed file systems (like Cephfs). My work is on tuning
btrfs with cephfs to span the entire network of computers I run[1].
Yes, cephfs' the hammer branch does' use RDMA or  RoCE, Rdma over Conformed
Ethernet, so infiband interfaces are not needed. Rdma as a concept will
eventually provide a mechanism for the DDR memory resources of a video
card to be use as system resources, imho. Many are working on this and
keeping their details private, for now.


> > ... does it use this new stuff anyway, do we need a specific USE-flag
> > enabled (I can't spot it, looking for something like "acc" or "rdma"
> > ), do we need specific CFLAGS .. ?

You have to dig a bit as most projects using this stuff are alpha or testing
stages. Here's one, ceph (hammer branche 0.94.x)  [2] 


> I can't speak for RDMA (can't find any mention of it in gcc) because
> that's an even more exotic thing than plain old accelerator support
> (unless you run infiniband at home...), but the flags are:

Rdma is often found in Rdma over Conformed Ethernet as the moniker for
language searches. RDMA is a concept, like DMA. RDMA will allow that
DDR5 memory on your graphics card to become part of the
compiler/computational general pool of  resources of your system. Vendors
might only do this on new platforms (like arm based servers). It's not
guaranteed that a method will be published for legacy hardware. That
sort of efforts is up to the FOSS/kernel folks.

>     -fopenmp
>     -foffload
>     -fopenacc

> However enabling them is as useful as having CFLAGS=-fopenmp currently. It
> changes __nothing__ unless an application has openmp annotations, and the
> ones that do should already provide a means of doing so in the build
> system.

Yes, this is all evolving as we speak, I never stated it was part of stable
applications. The advice was for one to consider when purchasing new
hardware, so a year or 2 from now they dont say they were not informed
of these fundamental changes to gcc and it's implications. In the past
special compilers, expensive hardware and lots of vendor trickery kept these
sorts of technologies from the masses of FOSS devs and users. Now this has
changed but it is going to be some effort and take a while for the greater
FOSS communities to learn how to leverages these new resources, which are a
'game changer'. Yes, right now these are the toys of the cluster folks, but
soon they'll be commonly used. Plan on replacing NFS with Ceph. Plan
on using distcc on a cluster [3].



> tldr: don't buy a dedicated gpu just because you read something on a
> mailing list ;)

Exactly correct. But if you are going to spend on new resources, be aware
that the gpu and ddr memory are now 'targets' for some very aggressive
development and all available via gcc-5. Apologies if what I wrote
seemed to be 'overly optimistic'. 


hth,
James

[1]
http://www.networkcomputing.com/storage/fibre-channel-really-is-dead/a/d-id/1320090

[2] https://community.mellanox.com/docs/DOC-2141

[3] https://github.com/mesos/mesos-distcc

[gentoo-user] Re: new computer : any advice ?

[james]

Jeremi Piotrowski <jeremi.piotrowski <at> gmail.com> writes:


> > AFAI understand gcc-5 should compile faster?
> > And generate faster code in some cases?

> No, and yes. Compilation is not affected in any way and runtime
> performance can only be improved _if_ this stuff is explicitly used within
> the code.

OpenAcc-2.0 is dated. Have you seen the latest draft of OpenAcc-2.5:: [1]?

Things are progressing at a brisk pace with openacc, imho. ymmv.

hth,
James

[1] http://www.openacc.org/content/openacc-25-draft-public-comment

Re: [gentoo-user] Re: new computer : any advice ?

[Fernando Rodriguez]

On Wednesday, September 09, 2015 9:52:55 PM james wrote:
> Jeremi Piotrowski <jeremi.piotrowski <at> gmail.com> writes:
> 
> > No, and yes. Compilation is not affected in any way and runtime
> > performance can only be improved _if_ this stuff is explicitly used within
> > the code.
> 
> Yes this is all new and a work in progress. I do not think it will be
> gcc-6 that makes the difference in a few years. But folks should be aware
> and look for codes that are accelerated via usage of GPU resources.
> Remember this all started about hardware purchase and future benefits.
> It's definitely not commodity usage atm.
> 
> 
> > Meaning you would feel a difference in no less then 5 years when gcc-6 is
> > widely used and accelerator support is not restricted to intel MIC and
> > nvidia gpus. James is getting a bit ahead of himself calling this a
> > "game changer" - yeaaaaah... not really right now.
> 
> It's not as restricted as you indicate amd, intel, nividia and others like
> arm (Mali and such) are working to support there hardware under the openacc
> code extension now in gcc-5. Granted the more powerful your GPU resources
> are the more they can contribute. This stuff use to only work with
> vendor supplied compilers and sdks, now it's finally available in gcc,
> albeit in it's infancy. Naturally it's going to take a while to 
> become mainstream useful; but that more like a year or 2, at most.

The value I see on that technology for desktop computing is that we get the 
GPUs for what they're made (graphics processing) but their resources go unused 
by most applications, not in buying powerful GPUs for the purpose of offloading 
general purpose code, if that's the goal you're better off investing in more 
general purpose cores that are more suited for the task.

To trully take advantage of the GPU the actual algorithms need to be rewritten 
to use features like SIMD and other advanced parallelization features, most 
desktop workloads don't lend themselves for that kind of parallelization. That 
is why despite similar predictions about how OpenMP-like parallel models would 
obsolete the current threads model since they where first proposed, it hasn't  
happened yet.

Even for the purpose of offloading general purpose code, it seems with all the 
limitations on OpanACC kernels few desktop applications can take advantage of 
it (and noticeably benefit from it) without major rewrites. Off the top of my 
head audio, video/graphics encoders, and a few other things that max out the 
cpu and can be broken into independent execution units.

-- 
Fernando Rodriguez

[gentoo-user] Re: new computer : any advice ?

[james]

Fernando Rodriguez <frodriguez.developer <at> outlook.com> writes:



> > albeit in it's infancy. Naturally it's going to take a while to 
> > become mainstream useful; but that more like a year or 2, at most.
> 
> The value I see on that technology for desktop computing is that we get the 
> GPUs for what they're made (graphics processing) but their resources go
unused 
> by most applications, not in buying powerful GPUs for the purpose of
offloading 
> general purpose code, if that's the goal you're better off investing in more 
> general purpose cores that are more suited for the task.


I think most folks when purchasing a workstation include a graphics
card on the list of items to include. So my suggestions where geared
towards informing folks about some of the new features of gcc that
may intice them to consider the graphics card resources in an
expanded vision of general resources for their workstation.


> To trully take advantage of the GPU the actual algorithms need to be
rewritten 
> to use features like SIMD and other advanced parallelization features, most 
> desktop workloads don't lend themselves for that kind of parallelization.

Not true if what openacc hopes to achived indeed does become a reality.
Currently, you are most correct. Things change; I'm an optimist because 
I see what is  occuring in embedded devices, arm64, and cluster codes.
ymmv.

> That 
> is why despite similar predictions about how OpenMP-like parallel models
would 
> obsolete the current threads model since they where first proposed, it
hasn't  
> happened yet.

Yes it's still new technology, controversial, just like systemd, clusters,
and Software Defined Networks.


> Even for the purpose of offloading general purpose code, it seems with all
the 
> limitations on OpanACC kernels few desktop applications can take advantage of 
> it (and noticeably benefit from it) without major rewrites. Off the top of my 
> head audio, video/graphics encoders, and a few other things that max out the 
> cpu and can be broken into independent execution units.


You are taking a very conservative view of things. Codes being worked
out now for clusters, will find their way to expand the use of the
video card resources, for general purpose things. Most of this will
occur as compiler enhancements, not rewriting by hand or modifying 
algorithmic designs of existing codes. Granted they are going to
mostly apply to multi-threaded application codes.


When folks buy new hardware, it is often a good time to look at what
is on the horizon for computers they use. All I have pointed out is
a very active area that benefits folks to review for themselves. I not
pushing expenditures of any kind on any hardware.

Caveat Emptor.



James

Re: [gentoo-user] Re: new computer : any advice ?

[Rich Freeman]

On Thu, Sep 10, 2015 at 8:20 AM, james <wireless@tampabay.rr.com> wrote:
> I think most folks when purchasing a workstation include a graphics
> card on the list of items to include. So my suggestions where geared
> towards informing folks about some of the new features of gcc that
> may intice them to consider the graphics card resources in an
> expanded vision of general resources for their workstation.

Sure, but keep in mind depreciation.

If all you need today is a $30 graphics card, then you probably should
just spend $30.  If you think that software will be able to use all
kinds of fancy features on a $300 graphics card in two years, you
should just spend $30 today, and then wait two years and buy the fancy
graphics card on clearance for $10.

It is pretty rare that it is a wise move to spend money today on
computer hardware that you don't have immediate plans to use.  The
only time it might make sense is if some kind of QA process means that
you're going to spend a lot more money on re-qualifying your system
after the upgrade than it would cost to just do it once and overspend
on hardware.  However, in general I'm not a big fan of those kinds of
QA practices in the first place.

-- 
Rich

Re: [gentoo-user] Re: new computer : any advice ?

[Gevisz]

On Thu, 10 Sep 2015 12:20:39 +0000 (UTC) james <wireless@tampabay.rr.com> wrote:

> Fernando Rodriguez <frodriguez.developer <at> outlook.com> writes:
> 
> > > albeit in it's infancy. Naturally it's going to take a while to 
> > > become mainstream useful; but that more like a year or 2, at most.
> > 
> > The value I see on that technology for desktop computing is that we
> > get the GPUs for what they're made (graphics processing) but their
> > resources go unused by most applications, not in buying powerful
> > GPUs for the purpose of offloading general purpose code, if that's
> > the goal you're better off investing in more general purpose cores
> > that are more suited for the task.

It is true.
 
> I think most folks when purchasing a workstation include a graphics
> card on the list of items to include. So my suggestions where geared
> towards informing folks about some of the new features of gcc that
> may intice them to consider the graphics card resources in an
> expanded vision of general resources for their workstation.
> 
> > To trully take advantage of the GPU the actual algorithms need to be
> > rewritten to use features like SIMD and other advanced parallelization
> > features, most desktop workloads don't lend themselves for that kind
> > of parallelization.

And it is also true.

> Not true if what openacc hopes to achived indeed does become a reality.

Hopes almost never becomes a reality.

> Currently, you are most correct.

Absolutely correct.

...

> 
> When folks buy new hardware, it is often a good time to look at what
> is on the horizon for computers they use.

I also considered "what is on the horizon" when bought a brand new
ATI Radeon R4770 graphic card about 6 years ago for computing purposes.

In half a year it was discovered that it has much worse performance than
ATI guys hoped for and, to improve it, they have to rewrite their proprietary
drive for this graphic card.

Instead of doing it, they just shamelessly dropped the support of the parallel
computing feature of this graphic card in all subsequent versions of their drive.

And as far as I know, no open source drive have ever supported the parallel
computing feature of this graphic card as well.

So, it was just a waste of money. Even more: I almost never worked at my assembled
almost 7 year-old 4-core AMD computer with this graphic card as for all other
purposes I prefer to work at my 10 year-old 2-core AMD computer with a very cheap
on-board video card. Just to avoid extra heating and aircraft noise produced by R4770.

So, Rich Freeman was absolutely right when he wrote in reply to your words above that

> If all you need today is a $30 graphics card, then you probably should
> just spend $30.  If you think that software will be able to use all
> kinds of fancy features on a $300 graphics card in two years, you
> should just spend $30 today, and then wait two years and buy the fancy
> graphics card on clearance for $10.

> It is pretty rare that it is a wise move to spend money today on
> computer hardware that you don't have immediate plans to use.

[gentoo-user] Re: new computer : any advice ?

[james]

Gevisz <gevisz <at> gmail.com> writes:

> on-board video card. Just to avoid extra heating and aircraft noise 
> produced by R4770.

Fanless video cards are wonderful. I have had many over the years but this
one is still my (silent) favorite::

01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI]
Cape Verde PRO [Radeon HD 7750 / R7 250E]


I did notice in a gentoo blog that openmp is a testing option for Clang-3.7
now? [1]

Try not to loose faith, we all have had bad experiences, but clustering,
distributed and systems aggregation codes are rapidly coalescing into
something wonderful, so ..... keep the faith...... bro.

;-)	

wwr,
James


[1] http://blog.cafarelli.fr/2015/09/testing-clang-3-7-0-openmp-support/

Re: [gentoo-user] Re: new computer : any advice ?

[Gevisz]

On Thu, 10 Sep 2015 21:12:37 +0000 (UTC) james <wireless@tampabay.rr.com> wrote:

> Gevisz <gevisz <at> gmail.com> writes:
> 
> > on-board video card. Just to avoid extra heating and aircraft noise 
> > produced by R4770.
> 
> Fanless video cards are wonderful. I have had many over the years but this
> one is still my (silent) favorite::
> 
> 01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI]
> Cape Verde PRO [Radeon HD 7750 / R7 250E]

Thank you for information.

Never say never, but according to my current mood
(which is unchanged for the last 5 years already :),
I will never buy ATI video card again.

And not because of noise and heating of ATI Radeon R4770
but because ATI shamelessly dropped support of their much
advertised parallel computing feature for this video card.

And I am not a fan of NVidia either.

Only small, cheap and fanless on-board video cards
to just manage the monitor!

At least for the next 12 or 15 years. :)
 
> I did notice in a gentoo blog that openmp is a testing option for Clang-3.7
> now? [1]
> 
> Try not to loose faith, we all have had bad experiences, but clustering,
> distributed and systems aggregation codes are rapidly coalescing into
> something wonderful, so ..... keep the faith...... bro.
> 
> ;-)	
> 
> wwr,
> James
> 
> 
> [1] http://blog.cafarelli.fr/2015/09/testing-clang-3-7-0-openmp-support/
> 
> 
> 
> 

[gentoo-user] Re: OT: GCC 5 Offloading

[Fernando Rodriguez]

On Thursday, September 10, 2015 12:20:39 PM james wrote:
> You are taking a very conservative view of things. Codes being worked
> out now for clusters, will find their way to expand the use of the
> video card resources, for general purpose things. Most of this will
> occur as compiler enhancements, not rewriting by hand or modifying 
> algorithmic designs of existing codes. Granted they are going to
> mostly apply to multi-threaded application codes.

Your being over-optimistic. It seems to me all they're hoping for is to define 
a standardized and portable high-level interface for programming accelerators. 
The ones that will benefit the most is the same applications that can benefit 
from lower level technologies like CUDA. Scientific/number crunching 
applications, some kinds of clustering, etc.

With no synchronization most existing multithreaded designs cannot benefit from 
it. And obviously code running on the accelerator cannot branch into the CPU, 
so no system or library calls. That leaves only purely number crunching loops. 
There's little of that on desktop and few of them can be fully optimized for 
parallelization. And to be worth the overhead of offloading the CPU needs to be 
maxed out. That leaves only the few applications I mentioned before.

I'm looking at Intel MICs[1] and those look a lot more promising though still 
of limited use for desktops. It uses OpenMP so it has a lot less restrictions 
than OpenACC (a few ebuilds in the tree can already benefit from it with minor 
patches) and you can even offload whole proccesses. You can even ssh to the MIC 
since it runs Linux. It's not for the average desktop but they're not too 
expensive either. It may be worth it for high-end gentoo workstation (you can 
offload compile jobs with distcc) and I got a project on the backburner that can 
benefit from it.

Do you know of any plans to enable offloading on the gentoo toolchain? I was 
able to build the offloading compiler using crossdev with a few hacks and wrote 
an ebuild for Intel's simulator[2]. I will work on enabling the host compiler 
tomorrow and may open a feature request and post patches once I get it 
working. The changes needed to enable it on the host are pretty trivial.

[1] https://software.intel.com/en-us/articles/intel-xeon-phi-coprocessor-codename-knights-corner

-- 
Fernando Rodriguez