Announcement

Collapse
No announcement yet.

AMD Athlon 5350 APU On Linux

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #61
    Originally posted by curaga View Post
    The AM1 socket is single-channel only?
    Yeah caught that a bit after posting and frankly that sucks. Still I like the board and processor combo, it is cheap enough to throw in a control panel to handle light data processing or MMI chores. It is certainly far better than most of the AToM offerings. Well maybe not this specific board, it will be interesting to see what the market delivers for this platform.

    Comment


    • #62
      Originally posted by tuke81 View Post
      Yeah, 450W psu for 25W soc is like a shooting fly with a tank(it could run easily dozen of them). There is no really good low power psus out there, maybe seasonic tfx psus like this: http://www.newegg.com/Product/Produc...82E16817151113 , same series has 250W model too. I would like to see some 100W-150w gold/platinium rated sfx/tfx psu, for this kind of stuff(I'm not so keen on picopsus).
      With all of this new low power hardware arriving the industry really needs a new power supply standard. Frankly it is hard to even justify the current bundle of wires approach to PC power supplies. The first thing to consider is a a single voltage input for the entire board with maybe a stand by line for waking up from low power modes. This could easily be handled by a four pin connector.

      My thinking is that 24VDC would be ideal as it works well with industrial and commercial vehicle systems. Considering the capabilities of modem power supply topologies, the 24 VDC could be nominal support a wide voltage input range. This would immediately eliminate the power supply problem for the low power boards as 24 VDC power supplies are extremely common. You can get them as wall warts, bench top, panel mount, chassis mount, open frame and a host of other styles in power ranges of a few watts to thousands of watts.

      Like the floppy, the CD drive and a bunch of other legacy hardware it is time for the industry to set a new standard for low power computers.

      Comment


      • #63
        Originally posted by Kivada View Post
        Yeah, but then you have the problem of finding a TFX case, so unless you have one of those Dell or HP buisness mini cases you are SOL unless you don't mind cutting a hole in the side of a standard case to allow you to plug it in.
        Companies like Corsair make both power supplies and computer cases. If they really wanted to, they could easily release a new array of low watt power supplies and cases made to fit them at the same time. Then, you wouldn't have an argument to make in regards to cases.

        Comment


        • #64
          Originally posted by mmstick View Post
          Companies like Corsair make both power supplies and computer cases. If they really wanted to, they could easily release a new array of low watt power supplies and cases made to fit them at the same time. Then, you wouldn't have an argument to make in regards to cases.
          Afaik corsair isn't psu oem, it uses different oems like seasonic, channel well, chicony or flextronics.

          Comment


          • #65
            Originally posted by tuke81 View Post
            Yeah, 450W psu for 25W soc is like a shooting fly with a tank(it could run easily dozen of them). There is no really good low power psus out there, maybe seasonic tfx psus like this: http://www.newegg.com/Product/Produc...82E16817151113 , same series has 250W model too. I would like to see some 100W-150w gold/platinium rated sfx/tfx psu, for this kind of stuff(I'm not so keen on picopsus).
            There are also thin mini itx motherboards which support laptop adapters which can reach 90% efficiency.

            Comment


            • #66
              Originally posted by wizard69 View Post
              With all of this new low power hardware arriving the industry really needs a new power supply standard. Frankly it is hard to even justify the current bundle of wires approach to PC power supplies. The first thing to consider is a a single voltage input for the entire board with maybe a stand by line for waking up from low power modes. This could easily be handled by a four pin connector.

              My thinking is that 24VDC would be ideal as it works well with industrial and commercial vehicle systems. Considering the capabilities of modem power supply topologies, the 24 VDC could be nominal support a wide voltage input range. This would immediately eliminate the power supply problem for the low power boards as 24 VDC power supplies are extremely common. You can get them as wall warts, bench top, panel mount, chassis mount, open frame and a host of other styles in power ranges of a few watts to thousands of watts.

              Like the floppy, the CD drive and a bunch of other legacy hardware it is time for the industry to set a new standard for low power computers.
              I have been saying this for years but nobody listens.

              Comment


              • #67
                Originally posted by JS987 View Post
                There are also thin mini itx motherboards which support laptop adapters which can reach 90% efficiency.
                True, like this(but it's like a build in pico psu):
                http://www.asrock.com/mb/AMD/AM1H-ITX/

                Comment


                • #68
                  Originally posted by przemoli View Post
                  But You can forget about reproductability.

                  Cause of implementation differences, cause of default settings differences, cause of implementation perf. differences. Etc.
                  How so? PTS is made to be reproducible, it uses the exact same x264 version no matter the host OS. Furthermore x264 uses its own asm routines, so the host compiler doesn't affect the result either.

                  Encoding perf obviously doesn't affect the generated file. If you encode from the same source, using the same x264 version, using the same settings, it's going to be comparable and reproducible if not bit-identical.

                  Comment


                  • #69
                    Any news on those wonky integer performance numbers? I have an E-450 running at 1.65 GHz and theoretically, the Athlon 5350 should be able to outperform it by a factor of 3-ish if perfect scaling is assumed (+20% IPC per core * 2x cores * 2.05 GHz /1.65 GHz = 1.2 * 2 * 1.2424... =~ 3). The single threaded theoretical performance gain alone should reach +50% fer chrissakes!

                    @Michael:

                    Would you be open to running both the E-350 and the Athlon 5350 using the performance governor to rule out any CPU frequency issues? Assuming the same disk I/O and RAM subsystem (preferably the 240GB OCZ Vertex 3 SSD and the 4GB DDR3-1600 MHz block), I would imagine that the Athlon would wipe the floor with the E-350 on e.g. the Linux kernel benchmark run tbh. If it doesn't, well...

                    Comment


                    • #70
                      Originally posted by mmstick View Post
                      Companies like Corsair make both power supplies and computer cases. If they really wanted to, they could easily release a new array of low watt power supplies and cases made to fit them at the same time. Then, you wouldn't have an argument to make in regards to cases.
                      Silverstone also makes cases, the aforementioned PSU has cables too short for some of their own ITX cases.

                      That and like mobos, not every case suits every usage type. What if I want a small but spacious file server? I could get n ITX board with 12x SATA ports like the ASRock C2750D4I in a case like the Fractal Node 304, a case you could easily fill with 12x 2.5" drives or 6x 3.5".

                      Comment

                      Working...
                      X