VOGONS


Socket-A AMD Athlon/Sempron system

Topic actions

First post, by gerwin

User metadata
Rank l33t
Rank
l33t

As requested by retro games 100, here a bit about the AMD system that I am typing this on. Never intended to be a retro or a high-end machine. Several components were just the best buy at a certain time. This PC is still used daily by me and my family, it is plenty fast if you keep it in order and use a not to heavy OS.

CPU: AMD Sempron 3000+ (12x166Mhz = 2000MHz) 512 kB L2 cache. 1.60 Volt (has a Zalman copper heatsink with a manually speed-regulated fan)
Mainboard: MSI KT4V 6712 (VIA KT400 chipset)
RAM: 1536MB DDR RAM @ 2x166MHz = 333MHz
Harddisk: WD 80GB IDE
Video: MSI Geforce 7600 GS, no fan
PSU: Chieftech 360 Watt max.
OS: Windows 2000 SP4 + DOS 7.10 (Earlier it ran several years with Windows 98SE)
Benchmark: Currently 13385 3DMarks'01

The AMD Sempron 3000+ CPU has the same Barton core as the AMD Athlon XP's, it could also have been called 'Athlon XP 2700+'. So it is just a marketing thing. The Sempron does have a slighlty lower voltage: 1.60V instead of the usual 1.65V. I don't know why the C.B.id screenshot says thorton core?. A must: Enable HALT state and STPGNT state (in your BIOS, or with 'Central Brain identifier' or with 'Cool-on'), it will decrease your Athlon's heat output dramatically, without any drawbacks. Most people advise to go for unlocked Athlons or rather Athlons XP-M(obile), yet the actual advantages over this Sempron are not that big.

The MSI 6712 mainboard: AGP-8X + 6 PCI slots. No ISA slots. Look at Prophase_j's KT133 system if you want an ISA slot. The KT400 chipset works well with 333MHz memory speed. It was also expected to support 400 MHz net memory speed, but I read it did not work well. 400MHz kinda worked on the later KT600 (I have one of these too) but it is not that interesting. The MSI 6712 is one of the last mainboards to support DDMA/NMI for legacy sound in dos. The mainboard has solid support for S3 'Suspend To Ram' standby state which is a great feature.

.

Attachments

  • sempron3000_b.gif
    Filename
    sempron3000_b.gif
    File size
    17.31 KiB
    Views
    7980 views
    File comment
    Central Brain identifier tab 2
    File license
    Fair use/fair dealing exception
  • sempron3000_a.gif
    Filename
    sempron3000_a.gif
    File size
    16.99 KiB
    Views
    7980 views
    File comment
    Central Brain identifier tab 1
    File license
    Fair use/fair dealing exception
  • athlon_cpus.jpg
    Filename
    athlon_cpus.jpg
    File size
    69.06 KiB
    Views
    7980 views
    File comment
    some AMD Athlon CPU's ranging 1100MHz to 2133MHz, the ones on the right are Barton core.
    File license
    Fair use/fair dealing exception
  • ms6712_specs.png
    Filename
    ms6712_specs.png
    File size
    77.74 KiB
    Views
    7980 views
    File comment
    Mainboard Specs
    File license
    Fair use/fair dealing exception
  • ms6712.jpg
    Filename
    ms6712.jpg
    File size
    160.82 KiB
    Views
    7980 views
    File comment
    Photo
    File license
    Fair use/fair dealing exception

Reply 4 of 22, by retro games 100

User metadata
Rank l33t
Rank
l33t

Awesome rig Gerwin! 😎 I note with interest, this sentence: "The MSI 6712 is one of the last mainboards to support DDMA/NMI for legacy sound in dos." That's something I must watch out for. (Also, when I look out for a "fast as possible, no ISA slot" mobo, I need to ensure it has win98 mobo chipset drivers.)

Reply 5 of 22, by fillosaurus

User metadata
Rank Member
Rank
Member

Hmm, only 13385 3DMarks'01 with a Geforce 7600 GS? My previous main rig was a socket 754 Sempron 2600+, Asus K8V-X mobo (VIA K8T800 chipset), 512 Mb DDR400, 80 Gb WD and a GeForce 6200. Around 10500 3DMarks'01, AFAIR.

Y2K box: AMD Athlon K75 (second generation slot A)@700, ASUS K7M motherboard, 256 MB SDRAM, ATI Radeon 7500+2xVoodoo2 in SLI, SB Live! 5.1, VIA USB 2.0 PCI card, 40 GB Seagate HDD.
WIP: external midi module based on NEC wavetable (Yamaha clone)

Reply 6 of 22, by retro games 100

User metadata
Rank l33t
Rank
l33t

@gerwin,

You said in your O.P. - "Video: MSI Geforce 7600 GS ... OS: <snip> Earlier it ran several years with Windows 98SE."

Does the 7600 have win98 drivers? I thought (perhaps incorrectly) that the last nVidia graphics card to have official win98 drivers was the 6xxxx series, such as the 6200 card for instance.

Reply 7 of 22, by gerwin

User metadata
Rank l33t
Rank
l33t

@swaaye
Concerning the errors in the CPU identification: Maybe in the past I did the 'wire-trick' on the AMD CPU to see if I could unlock something, and left it like that.

@retro games 100
Yes, The 7600 came after I had replaced windows 98 with windows 2000. The earlier Geforce 6600 had given up entirely, before that there was the MX440 again, and a Geforce 2 Ti with a noisy fan.

@fillosaurus
IIRC a Tualatin P3 with a geforce 4 should already come close to 10000 marks, so 13300 is not to impressive indeed. I disabled some things in the tray, and disabled 2xAA and set 'monitor TFT scaling', also I disabled CPU Halt/stpgnt. Now I get 14462 marks '01.

Video card specs:

MSI AGP 256MB NVD 7600GS - Fanless NX7600GS-TD256Z […]
Show full quote

MSI AGP 256MB NVD 7600GS - Fanless
NX7600GS-TD256Z

SPECS
Nvidia GeForce 7600 GS ( NV4B/G73 revision A2 (12pp,5vp) )
DDR2 SDRAM, 128 bit, 400 MHz (800Mhz effective)
RAMDAC 400 MHz

CONNECTORS
AGP 4x/8x, AGP revision 3.0
1 x DVI-I
1 x D-Sub
S-video out

SUPPORTS
DirectX version 9.0c
HDTV Ready
OpenGL version 2.0
Shader model version 3.0
Maximum resolution 2048x1536

MINIMUM SYSTEM REQUIREMENTS
Intel Pentium IV / AMD K8
RAM 128 MB

Reply 8 of 22, by leileilol

User metadata
Rank l33t++
Rank
l33t++
fillosaurus wrote:

Hmm, only 13385 3DMarks'01 with a Geforce 7600 GS? My previous main rig was a socket 754 Sempron 2600+, Asus K8V-X mobo (VIA K8T800 chipset), 512 Mb DDR400, 80 Gb WD and a GeForce 6200. Around 10500 3DMarks'01, AFAIR.

Roughly ditto but a x850, got 18500 3dmarks01

Sempron64s can really be pushed to great power (even the 2600s), it's one hell of a budget processor.

apsosig.png
long live PCem

Reply 9 of 22, by fillosaurus

User metadata
Rank Member
Rank
Member

Yea, my Sempron 2600+ was a nice beastie, but I think I had the older 32bit version. AFAIK, his current owner takes good care of it (he's a friend with a lower budget than mine, so when I got my current rig I said "Ama gonna make you an offer you can't refuse", and he didn't refused 😁).

Y2K box: AMD Athlon K75 (second generation slot A)@700, ASUS K7M motherboard, 256 MB SDRAM, ATI Radeon 7500+2xVoodoo2 in SLI, SB Live! 5.1, VIA USB 2.0 PCI card, 40 GB Seagate HDD.
WIP: external midi module based on NEC wavetable (Yamaha clone)

Reply 10 of 22, by gerwin

User metadata
Rank l33t
Rank
l33t

I got a little problem with this KT400 system

Crystal diskmark 2.2 Seq Read test:
KT400 with WD800JB IDE HDD: 28 MB/s
KT600 with WD800JB IDE HDD: 58 MB/s
BIOS settings are optimal for all I know, It was tested on a partition without much fragmentation.

Now when these identical harddisks were both new I had the same problem, but after some hours the KT400 began to give 55 MB/s results, like the KT600 does reliably. So I was satisfied again. But now the results are again halved for unknown reason.
What can I do?

--> ISA Soundcard Overview // Doom MBF 2.04 // SetMul

Reply 12 of 22, by swaaye

User metadata
Rank l33t++
Rank
l33t++

I was looking over some old mobo reviews awhile ago and discovered that VIA disk controllers basically suck. It's something that I've suspected for years and looking up review numbers showed it to be very true.

If that mobo has a separate IDE RAID controller, try it instead. The Highpoint chips are very fast, for example. Just remember that you need to install the driver in Windows before you try to boot on it.

I wouldn't be surprised if performance varies from board to board due to BIOS differences. Maybe board quality too.

Reply 13 of 22, by Tetrium

User metadata
Rank l33t++
Rank
l33t++
swaaye wrote:

I was looking over some old mobo reviews awhile ago and discovered that VIA disk controllers basically suck. It's something that I've suspected for years and looking up review numbers showed it to be very true.

If that mobo has a separate IDE RAID controller, try it instead. The Highpoint chips are very fast, for example. Just remember that you need to install the driver in Windows before you try to boot on it.

I wouldn't be surprised if performance varies from board to board due to BIOS differences. Maybe board quality too.

I wasn't aware of that. So far I've been pretty satisfied with VIA chipsets even though I know it may not be the fastest on the block.

Couldn't it be the harddrive btw?

Reply 14 of 22, by Old Thrashbarg

User metadata
Rank Oldbie
Rank
Oldbie

I've found that Via disk controllers tend to work fine as long as you don't install the Via drivers for them... I always stick with the MS built-in drivers, and have never had any issues.

Reply 15 of 22, by swaaye

User metadata
Rank l33t++
Rank
l33t++

Yeah there's not much point in installing drivers for PATA controllers that work as generic IDE hardware. Regardless of manufacturer. I've never seen an improvement in performance. On the other hand, I have definitely found problems with 3rd party IDE drivers.

Regarding VIA performance though see:
http://techreport.com/articles.x/4800/8
http://techreport.com/articles.x/3590/8
http://techreport.com/articles.x/3802/7 (watch HPT372 dust ICH4!)

The only reason I looked for more info on this was because over the years I've noticed that VIA systems seemed sluggish to me compared to contemporary NVIDIA/Intel chipsets. One of the more potent moments of this was with two notebooks, one with ICH6M and the other with VT8235, both using the same 7200RPM 2.5" drive, and the ICH6M being clearly more responsive.

At work a few months ago I resurrected a XP sys based on an old ALI Magick 1-based board. It has a HPT372 chip. I was having problems with the ALI IDE so decided to try out the HPT372. Not only has it been trouble free for months now but it is tangibly faster. I used to distrust those HPT chips, since I had problems with some 440BX mobos with the older HPT36x chip, but this has changed my stance on them.

If there's a RAID chip on a mobo, might as well try it out, especially if you have a non-Intel chipset.

Reply 16 of 22, by gerwin

User metadata
Rank l33t
Rank
l33t

Thanks for the response, I have found another flatcable (yes one with 80 wires) and will first try if that improves the transfer speed on the KT400.
Sorry for the delay, but a router and netbook (1000HE) just came in and I just could not resist to set them up first.

--> ISA Soundcard Overview // Doom MBF 2.04 // SetMul

Reply 17 of 22, by gerwin

User metadata
Rank l33t
Rank
l33t
gerwin wrote:

Thanks for the response, I have found another flatcable (yes one with 80 wires) and will first try if that improves the transfer speed on the KT400.

Tried the new cable, but it doesn't make any difference. 🙁

--> ISA Soundcard Overview // Doom MBF 2.04 // SetMul

Reply 18 of 22, by ux-3

User metadata
Rank Oldbie
Rank
Oldbie

What about DMA? You know that windows will shut the fast access down eventually. I had this a few times when I was trying to safe a disk which was damaged. Too many read errors make windows use pio mode again - for good! Check in the hardware manager

Retro PC warning: The things you own end up owning you.

Reply 19 of 22, by gerwin

User metadata
Rank l33t
Rank
l33t
ux-3 wrote:

What about DMA? You know that windows will shut the fast access down eventually. I had this a few times when I was trying to safe a disk which was damaged. Too many read errors make windows use pio mode again - for good! Check in the hardware manager

Ah, there I found it, but it says "Ultra DMA" already... DVD drive at secundairy IDE was set at PIO though, so I set it to Ultra DMA as well.
Next I installed VIA_HyperionPro_V524A chipset drivers. No difference in Crystal Diskmark: averaging 27.

--> ISA Soundcard Overview // Doom MBF 2.04 // SetMul