GA-5AA (rev 2.2) owner opinions?

Discussion relating to Socket 7 hardware.
Post Reply
sylvan2626
Senior K6'er
Posts: 139
Joined: Tue Apr 05, 2005 3:25 pm
Location: Florida, USA

GA-5AA (rev 2.2) owner opinions?

Post by sylvan2626 »

Any opinions on this mobo out there?

I saw on http://SPAM/ that tazwegion is/was an owner. I'm looking at a replacement mobo.
sylvan2626
Senior K6'er
Posts: 139
Joined: Tue Apr 05, 2005 3:25 pm
Location: Florida, USA

Post by sylvan2626 »

Anybody ever used one of these boards, any version?
User avatar
tazwegion
Veteran K6'er
Posts: 349
Joined: Sat Jun 21, 2003 11:13 am
Location: Victoria, Australia

Post by tazwegion »

Yeah... the GA-5AA is the AT version, and the GA-5AX the ATX respectively ;)

I originally sourced one because of three reasons...

1) I like Gigabyte build quality/reliability (and ease of parts replacement)

2) I needed a baby AT platform for my faithfull AT mid tower

3) The FSB settings namely 66, 75, 83, 95, 100 to 140 in 5Mhz increments

Interestingly enough, the 5AA also has an ATX power connection, I'm not overly enthusiastic about the ALi chipset, prefering the VIA apollo of the same time period, though that doesn't detract from the board in anyway it's merely a personal bias, it was up until it's (post-repair) demise a solid little performer and one of my many distributed computing platforms, whilst I haven't really attempted to overclock this mobo yet, it's nice to have more options ;)

For what it's worth mine is the 3.2 revision and it successfully supported a K62-550 ;)
Image
Super_Relay
Site Admin
Posts: 215
Joined: Sat Jun 02, 2001 5:53 pm
Location: Australia

Post by Super_Relay »

I have one of these too. I'm not sure what revision it is (i would look it up but its actually running in a machine my brother is using right now).

It ran my k63+ 450 at 575 with a fsb of 115mhz

i had a geforce 2mx in it and it ran great. the only thing was i had to force it to agp1x or else it would get flakey on me but after putting it at 1x it was rock solid.
Jim
K6'er Elite
Posts: 1745
Joined: Wed Jan 21, 2004 7:10 pm
Location: Toronto

Post by Jim »

Very curious Taz; why do you favour the VIA chipset? From what little I know the ALI is: a) easier to work with; b) has better memory bandwidth; and c) generates better numbers in the Everest benchmarks. Is there some other reason the VIA is better that I don't know about? Or is my experience with the VIA limited by the fact that I have only had DFIs and FICs.
Superpuppy 3
K6-3+ 450 ACZ (6x100)
DFI K6BV3+/66 Rev B2 (2 Meg) w/ 2x28mm Chipset Fans
2x256 Meg PC 133 Hynix SDRAM
1x 20G Maxtor (7200)
2x 80G Maxtor (7200) Ducted w/ 2x486 Fans Mount
52/24/52/16 LG CDR/RW/DVD
8/4/3/12/24/16/32 LG Super Multi
ATI 9000 aiw Radeon AGP
SB Audigy 1 MP3 Sound
CMD 649 IDE Controller
NEC USB 2 Card
User avatar
tazwegion
Veteran K6'er
Posts: 349
Joined: Sat Jun 21, 2003 11:13 am
Location: Victoria, Australia

Post by tazwegion »

Well Jim, I like a multitude of options/settings for example... the ability to set either synchronus OR asynchronus memory speeds (allowing for use of PC66 or 72 pin RAM), the MVP3 chipset will also support a larger L2 cache of 2Mb whereas the ALi can only support a MAX of 1Mb, and lastly ALL the MVP3 chipset m/boards I've ever owned have supported CPU temperature/voltage monitoring, this was NOT the case with the little GA-5AA ;)

As far as perfomance goes, I would assume that... "He who haveth' the largest L2 is triumphant!" :twisted:
Jim
K6'er Elite
Posts: 1745
Joined: Wed Jan 21, 2004 7:10 pm
Location: Toronto

Post by Jim »

I think that you know more about this than I do, so please don't take my comments and questions the wrong way. I thought that "Tag Ram" had something to do with how much Ram could be cached; and that the ALI was better in that regard than the MVP3. Something to do with an unused "Dirty Bit" that allowed the ALI to cache more Ram per k of cache than the MVP3.

I also thought "the more cache the better"; but I have seen posts stating that the K6-3+ can cache 512 Meg of Ram by itself; and that L3 cache had very little effect on performance.

I know that my ASUS P5A-B Rev 104s, (ALI), operate in "Write Back" mode, whereas my DFI K6BV3+/66s, (MVP3), operate in "Write Thru" mode; and changing them to "Write Back" gives a huge hit in performance as the amount of cachable Ram is halved. I have seen the numbers in "Everest" while monkeying around with WPCredit.

There is also a post in here by Kyle Brant saying that the ALI "is a better memory speed performer"; and I have heard from someone else with respect of the DFI that the 1 Meg board is better than the 2 Meg because its FSB can be clocked higher, (an extra 10 or 12 Mhz), without having to disable the on board cache.

I know that the P5A-B supports temperature and voltage monitoring; and even gives readouts of the actual values, (which the DFI does not for voltages), so I don't think that is a chipset factor, so much as a mobo design factor.

I didn't know the ALI was limited to 1 Meg cache; and had thought that the ideal Super 7 would be an ALI board w/ 2Meg cache onboard, ATA/66, and 4 PCI slots. (If there were such a board).

Any comments , and explanations, (especially rebuttals of false information), would be appreciated.

Thanks.
Superpuppy 3
K6-3+ 450 ACZ (6x100)
DFI K6BV3+/66 Rev B2 (2 Meg) w/ 2x28mm Chipset Fans
2x256 Meg PC 133 Hynix SDRAM
1x 20G Maxtor (7200)
2x 80G Maxtor (7200) Ducted w/ 2x486 Fans Mount
52/24/52/16 LG CDR/RW/DVD
8/4/3/12/24/16/32 LG Super Multi
ATI 9000 aiw Radeon AGP
SB Audigy 1 MP3 Sound
CMD 649 IDE Controller
NEC USB 2 Card
Guest

Post by Guest »

Yeah I've read that post/thread by Kyle, but I was coming from the point of view of a K62 not a plus version with on-die L2, also I stated it was my personal bias, I prefer the ability to run various type of RAM, and like the option of setting the memory speeds independant of the FSB/CLK speeds ;)

I've read of this TAG ram (which allows a cacheable area independant of the L2 size) but for various reasons didn't always work as they were designed, it would also be worth pointing out that the ALi chipset was the first on the market for supporting 100Mhz FSB in SS7 platforms (with MVP3 came later), perhaps the chipset was initially rushed one can only assume, but the many revisions to ALi platforms would seem to imply this fact. :wink:

Image

Image

More information @ ALi Vs. MVP3 - Hardware Upgrade

Beaten 5 out of 6 tests... my money's still on the VIA chipset (especially the Aopen & Epox platforms) still the moral to this story is... DON'T buy a LuckyStar m/board eh? :lol: :p
Jim
K6'er Elite
Posts: 1745
Joined: Wed Jan 21, 2004 7:10 pm
Location: Toronto

Post by Jim »

From that, it seems like the Epox is the way to go. No mention of DFI unfortunately, so I don't know where they fit in the scheme of things. Wish I had joined here sooner when Mr Toastz was still around. Do you have any comment on the idea that it is better to run a 1 Meg board at 112 FSB than a 2 meg board at 103?
Superpuppy 3
K6-3+ 450 ACZ (6x100)
DFI K6BV3+/66 Rev B2 (2 Meg) w/ 2x28mm Chipset Fans
2x256 Meg PC 133 Hynix SDRAM
1x 20G Maxtor (7200)
2x 80G Maxtor (7200) Ducted w/ 2x486 Fans Mount
52/24/52/16 LG CDR/RW/DVD
8/4/3/12/24/16/32 LG Super Multi
ATI 9000 aiw Radeon AGP
SB Audigy 1 MP3 Sound
CMD 649 IDE Controller
NEC USB 2 Card
User avatar
tazwegion
Veteran K6'er
Posts: 349
Joined: Sat Jun 21, 2003 11:13 am
Location: Victoria, Australia

Post by tazwegion »

Well, to be honest Jim, I wouldn't have the foggiest!

In theory most, if not all performance gains with the K6-X processor are to be had by upping the FSB rates on the platform, however this is totally dependant on the quality/speed of the L2 cache & memory installed, hence the old adage rings true, "you get what you pay for..." and this is best displayed by the poor performance of the EL-cheapo Lucky Star MVP3 m/board(s) in that particular review ;)


I guess the bottom line is... I'll have to go and personally run a 5 suite test on the 3 SS7 motherboards (Hmmmm... maybe I'll include the TXpro2 also :P) currently gracing my Retro collection, then post the results ;)

Sisoft Sandra Arithmetic benchmark,
Sisoft Sandra Multimedia benchmark,
CPUmark99
3Dmark2000
+
highest stable OC


Any particular test/benchmark you'd like to see?
Jim
K6'er Elite
Posts: 1745
Joined: Wed Jan 21, 2004 7:10 pm
Location: Toronto

Post by Jim »

Sandra 2004 or 2005, got em both. (mainly use 2004 because that gives me a comparison with old benchmarks made with my old setup) Also "Everest" since I use that. Never have had the others. Hope it isn't asking too much; and thanks. "Q" : Do You have a DFI?

I have found that I get the biggest performance gains so far from enabling "Write Allocate", setting the "DRAM Start Cycle" at "Start with cache at 66Mhz" as opposed to "Start after cache at 100Mhz", setting the "DRAM Read Pipeline Cycle" to "on", setting the "Cache Read pipline Cycle" to "on", and setting the "Cache Write Pipeline Cycle" to "on".

Additional gains can be had by enabling "4 Bank Interleave", setting "Backoff CPU till L2 Fill" to "wait", setting agressive Ram timings, and setting "PCI Master Directly Accesses DRAM if in GART Range" to "enabled".

There are other settings that give speed gains, (including 1 reserved setting that I changed and benched by accident !! - set that sucker back where it belonged and haven't touched it since), but I have found that I am losing stability even without them, since I installed the "VIA Latency Patch" and the "Unofficial Win98 Service Pack" on the 98 boot on my machine.

The XP side does not have those two, so I can still use a number of other tweaks there, and it runs a bit faster than the 98 boot as a result. Everest 299 vs Everest 287 Memory Read. More stable too.

So far I have not been able to get CPUCool to work in attempting to OC the FSB. Need advise there.

Anyway, thanks for your time.
Superpuppy 3
K6-3+ 450 ACZ (6x100)
DFI K6BV3+/66 Rev B2 (2 Meg) w/ 2x28mm Chipset Fans
2x256 Meg PC 133 Hynix SDRAM
1x 20G Maxtor (7200)
2x 80G Maxtor (7200) Ducted w/ 2x486 Fans Mount
52/24/52/16 LG CDR/RW/DVD
8/4/3/12/24/16/32 LG Super Multi
ATI 9000 aiw Radeon AGP
SB Audigy 1 MP3 Sound
CMD 649 IDE Controller
NEC USB 2 Card
User avatar
tazwegion
Veteran K6'er
Posts: 349
Joined: Sat Jun 21, 2003 11:13 am
Location: Victoria, Australia

Post by tazwegion »

No sorry I don't have a DFI (not a market heavy weight downunder), so all I can offer you is the 3-way shoot-out between the MVP3c2, AX59pro & GA-5AA, I was going to add a TXproII platform to the mix but it's lack of AGP would be an unfair appraisal in relation to 3Dmark 2000 ;)

UNLESS...
4Mb AGP x2 Vcard (66Mhz bus = 528 Mb/s)
16Mb PCI Vcard (33Mhz bus = 133Mb/s)


could be considered a fair match, ie. with the additional memory of the PCI supplimenting the lower bus speed & transfer rate? :?

Anyway... I've finished collating the results (18 screen shots :o), but to save space I'll have to 'cut and paste' 3-into-1 results, hopefully I should have it done/posted later today, though I might start a new thread as we've kinda hijacked this one eh? :P
Jim
K6'er Elite
Posts: 1745
Joined: Wed Jan 21, 2004 7:10 pm
Location: Toronto

Post by Jim »

Blame me for that.
Superpuppy 3
K6-3+ 450 ACZ (6x100)
DFI K6BV3+/66 Rev B2 (2 Meg) w/ 2x28mm Chipset Fans
2x256 Meg PC 133 Hynix SDRAM
1x 20G Maxtor (7200)
2x 80G Maxtor (7200) Ducted w/ 2x486 Fans Mount
52/24/52/16 LG CDR/RW/DVD
8/4/3/12/24/16/32 LG Super Multi
ATI 9000 aiw Radeon AGP
SB Audigy 1 MP3 Sound
CMD 649 IDE Controller
NEC USB 2 Card
Post Reply