VOGONS


Reply 40 of 56, by Windows98_guy

User metadata
Rank Newbie
Rank
Newbie
Tetrium wrote on 2022-02-03, 13:43:
ODwilly wrote on 2022-02-03, 13:41:

I wonder if a recap would help with stability on this board if you can't get a Windows OS stable with a 100% compatible CPU.

Back when I was working a lot with these boards, it didn't seem to be the caps (these boards were way younger then they are now).
These boards were finicky right out of the box, but still it might not even be a bad idea!

Do you remember how finicky were those motherboards back in the day?

Recently my ASUS A7V (also revison 1.02) has had 2 different problems at the same time. One, the SSD running through Sata to IDE adapter stopped being detected by the board. And it wasn't the problem with the adapter or the disk, because i tried different adapters and a Sata HDD and it still wasn't detected. However it did detect the SD Card to IDE adapter and about a week later, when i put the mb to a test bench for some tests, the SSD just started to work again - it detected it again!
Two, any PS2 keyboard i use, locks the computer and i get horrible PC speaker noises for a few seconds. Sometimes it even gives me no keyboard or keyboard error in the post screen.

Are these any of the common problems with this motherboard or are they completly new problems?

Reply 41 of 56, by 386SX

User metadata
Rank l33t
Rank
l33t

Many here has more experience of me on this mainboard serie but from tests in their past and in these times, it doesn't seem a myth the incredibly strange long list of things that can or cannot work on this board. Beside any SATA adapters powered by the PSU that I'd not use anymore on anything, even changing a single variable after having installed or configured the system, can result in crash or whatever. That include cpu version, memory version or mix of them, the first AGP or PCI card used before the second card etc.. the bios version itself (sometimes newer bios seems totally needed other times an older version helped..), AGP speed config, etc..

Probably to the old list of things not working or working nowdays we add weaker capacitors to the problem I'd not be surprised.

Reply 42 of 56, by Windows98_guy

User metadata
Rank Newbie
Rank
Newbie
386SX wrote on 2022-10-22, 16:45:

Many here has more experience of me on this mainboard serie but from tests in their past and in these times, it doesn't seem a myth the incredibly strange long list of things that can or cannot work on this board. Beside any SATA adapters powered by the PSU that I'd not use anymore on anything, even changing a single variable after having installed or configured the system, can result in crash or whatever. That include cpu version, memory version or mix of them, the first AGP or PCI card used before the second card etc.. the bios version itself (sometimes newer bios seems totally needed other times an older version helped..), AGP speed config, etc..

Probably to the old list of things not working or working nowdays we add weaker capacitors to the problem I'd not be surprised.

Well, thanks. I guess I could replace the capacitors, since they are 20+ years old, but i will wait and try to get a measuring tool, just to see how bad they are.

Reply 43 of 56, by rasz_pl

User metadata
Rank l33t
Rank
l33t

Im gonna guess caps will be near perfect
I had many (probably sold >20 systems) A7V133 and few A7V, the only issues I can recall would be with sensitive OC menu (jumperfree)

Open Source AT&T Globalyst/NCR/FIC 486-GAC-2 proprietary Cache Module reproduction

Reply 44 of 56, by karakarga

User metadata
Rank Newbie
Rank
Newbie

Windows98_guy,
"One, the SSD running through Sata to IDE adapter stopped being detected...." At Windows 98 there is no SSD support. Your SSD very quickly looses it's health. Windows 98 has it's ramdom writes to hard drives. I mean, Windows 98 all the time writes on disk. Writings are limited to SSD drives, they become aged very fast. Do not use SSD's unless sacrificing them.

I have one Asus K7V, I have repaired it, bought not working condition. I put a Athlon XP 800 processor. It works not bad actually. Asus site has a bios but it is not correct! "https://www.asus.com/supportonly/a7v/helpdesk_download/" They have mixed their database, long time passed. The size of it is 512kb but, this board's bios is only 256kb. A7V0206AS is the wrong one. A7V_101201a_beta is the correct bios. You can download from here: "https://drivers.eu/Mainboards/ASUS/A7V%28MB%29" and CPU support is as follows:

Processor -------------- PCB Bios
Athlon 1000 (200 MHz FSB) ALL ALL
Athlon 1100 (200 MHz FSB) ALL ALL
Athlon 1200 (200 MHz FSB) ALL ALL
Athlon 1300 (200 MHz FSB) ALL ALL
Athlon 1400 (200 MHz FSB) ALL 1008
Athlon 700 ALL ALL
Athlon 750 ALL ALL
Athlon 800 ALL ALL
Athlon 850 ALL ALL
Athlon 900 ALL ALL
Athlon 950 ALL ALL
Duron 1000 (Model 7)(Morgan) 1.05. 1008
Duron 1100 (Model 7)(Morgan) 1.05. 1008
Duron 1200 (Model 7)(Morgan) 1.05. 1008
Duron 1300 (Model 7)(Morgan) 1.05. 1008
Duron 600 (Model 3) ALL ALL
Duron 650 (Model 3) ALL ALL
Duron 700 (Model 3) ALL ALL
Duron 750 (Model 3) ALL ALL
Duron 800 (Model 3) ALL ALL
Duron 850 (Model 3) ALL ALL
Duron 900 (Model 3) ALL ALL
Duron 900 (Model 7)(Morgan) 1.05. 1008
Duron 950 (Model 3) ALL ALL
Duron 950 (Model 7)(Morgan) 1.05. 1008

The list is from A7PRO version, without Promise controller based model. "https://www.asus.com/supportonly/a7pro/helpdesk_cpu/" So, if your PCB revision is 1.02 (I also have the same revision) you can not use PCB 1.05 supported processors like Duron 1300.

Last edited by karakarga on 2022-12-02, 07:56. Edited 2 times in total.

Reply 47 of 56, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

Like I said on the first page - you can try bunch of unsupported CPUs and something most likely will boot up.

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 48 of 56, by agent_x007

User metadata
Rank Oldbie
Rank
Oldbie
karakarga wrote on 2022-12-01, 22:51:

At Windows 98 there is no SSD support. Your SSD very quickly looses it's health. Windows 98 has it's ramdom writes to hard drives. I mean, Windows 98 all the time writes on disk. Writings are limited to SSD drives, they become aged very fast. Do not use SSD's unless sacrificing them.

Modern SSDs have build-in garbage collection and TRIM (NAND controller does this automaticly), so they won't "die quick death" you describe here.
Yes, OS supports helps, but it's not needed unless you write gigabytes worth of data through a drive... each day (or you keep it at 90%+ capacity used state for prelong periods of time).
With 20-30% free space on modern SSD drive, you should be fine on longevity under Win98.

If you are still worried about it, just use Intel's older SSDs that have manual TRIM trigger option in Intel's app. You will need to dual boot with WinXP to trigger it, but it can be done.

Another option is a plain Compact Flash card (or Industrial version, for extended longevity).

157143230295.png

Reply 49 of 56, by rasz_pl

User metadata
Rank l33t
Rank
l33t

Build in garbage collection doesnt do anything if drive is full. Drive is considered full as soon as every sector has been written once since last Secure Erase. TRIM doesnt do anything if OS doesnt send deleted ranges back.
You can find a lot of speculation about magic garbage collection able to snoop on filesystem access patterns, but its all rumors. No product exists on the market with official magic trimless garbage collection of unused space on FAT/NTFS.
On XP/98 your best bet is running FAT32 and booting https://archive.org/details/trim_20190926 once a week.

Open Source AT&T Globalyst/NCR/FIC 486-GAC-2 proprietary Cache Module reproduction

Reply 50 of 56, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

Garbage collection was a thing even on now vintage SSDs, so it's not an argument towards modern SSDs. But most modern SSDs use SLC caching, which is especially beneficial for old systems.

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 51 of 56, by alvaro84

User metadata
Rank Member
Rank
Member

* by my thinking - I just don't see a way for an SSD to know which sectors are allocated or free, if they don't have their own file system driver in which case TRIM wouldn't even be a thing but the SSDs themselves would come with warnings in huge red letters how to partition them and what file systems they can interpret.

So, I think they have absolutely no idea what's "garbage" they are allowed to collect.

Shame on us, doomed from the start
May God have mercy on our dirty little hearts

Reply 52 of 56, by TrashPanda

User metadata
Rank l33t
Rank
l33t

We back on this old nugget again.

Dont fully partition your SSD drives, leave at least 10% of its volume as free unpartitioned space and the controller will handle garbage collection and allocation perfectly fine without a Trim aware OS. Modern SSDs have their own tracking of everything they need to handle the collection and allocation of cells, problems occur when users use 100% of the free space on the SSD and dont leave enough unpartitioned free space for the garbage collection systems to do their work.

Modern controllers are pretty much their own ASIC and the main OS simply communicates to them what data it wants, the drive handles everything else. (By modern Im talking about at least drives from the last 4-5 years)

The truth most people seem to ignore is that any modern SSD will easily outlast the retro system you are going to use it on, even without Trim or garbage collection, they have such huge read/write capacity that no retro rig is ever going to hit that limit.

Heck even modern systems wont hit that limit.

Reply 53 of 56, by rasz_pl

User metadata
Rank l33t
Rank
l33t

>Dont fully partition your SSD drives, leave at least 10% of its volume as free unpartitioned space

this will work only if nobody did anything with that drive before you, otherwise first step should be secure erase (lets controller know all logical sectors can be unlinked from data) and then partition

>any modern SSD will easily outlast the retro system you are going to use it on

took me a ~year to kill 32GB ADATA mSATA SSD in XP. Well, not really kill per se, wear out. It can write around 1GB (dynamic SLC buffer size?) at full speed and then BAM drops to 800KB/s for another couple megabytes to finally start hovering around 100 Bytes/second write speed 😀

Open Source AT&T Globalyst/NCR/FIC 486-GAC-2 proprietary Cache Module reproduction

Reply 54 of 56, by TrashPanda

User metadata
Rank l33t
Rank
l33t
rasz_pl wrote on 2022-12-04, 17:00:
>Dont fully partition your SSD drives, leave at least 10% of its volume as free unpartitioned space […]
Show full quote

>Dont fully partition your SSD drives, leave at least 10% of its volume as free unpartitioned space

this will work only if nobody did anything with that drive before you, otherwise first step should be secure erase (lets controller know all logical sectors can be unlinked from data) and then partition

>any modern SSD will easily outlast the retro system you are going to use it on

took me a ~year to kill 32GB ADATA mSATA SSD in XP. Well, not really kill per se, wear out. It can write around 1GB (dynamic SLC buffer size?) at full speed and then BAM drops to 800KB/s for another couple megabytes to finally start hovering around 100 Bytes/second write speed 😀

Thats not worn out .. that your drive hitting the Dram cache wall ..actually sounds more like the drive doesn't have any Dram cache outside of a small SLC one . .in other words its a cheap SSD using QLC Nand that would do exactly the same under Windows 10/11. Once your SLC cache is full the drive will drop its write speed drastically as it cannot clear the cache faster than it can write data to the Nand, there is nothing you can do to prevent this with Dramless SSDs, I have a few myself and they should be only used for storage and not as boot drives.

Wearing out the NAND would take several hundreds of terabytes and would take more than a year of constant 24/7 writes to achieve, even data centers take years to wear out SSDs and will generally replace them before they hit 80% capacity. You should read up on how SSD controllers handle worn out Nand cells.

So unless you are doing hundreds of gigabytes of writes per day on your retro rig I doubt you will ever kill a SSD with it.

Reply 55 of 56, by rasz_pl

User metadata
Rank l33t
Rank
l33t
TrashPanda wrote on 2022-12-04, 17:16:

Thats not worn out .. that your drive hitting the Dram cache wall ..actually sounds more like the drive doesn't have any Dram cache outside of a small SLC one . .in other words its a cheap SSD using QLC Nand that would do exactly the same under Windows 10/11. Once your SLC cache is full the drive will drop its write speed drastically as it cannot clear the cache faster than it can write data to the Nand, there is nothing you can do to prevent this with Dramless SSDs, I have a few myself and they should be only used for storage and not as boot drives.

its worn out all right. For the first year it did (from old crystaldiskmark log file from 2014/08/09):
Sequential Read : 225.239 MB/s
Sequential Write : 39.410 MB/s
Random Read 512KB : 215.342 MB/s
Random Write 512KB : 39.472 MB/s
Random Read 4KB (QD=1) : 30.953 MB/s [ 7557.0 IOPS]
Random Write 4KB (QD=1) : 39.629 MB/s [ 9675.2 IOPS]
Random Read 4KB (QD=32) : 49.314 MB/s [ 12039.5 IOPS]
Random Write 4KB (QD=32) : 38.827 MB/s [ 9479.2 IOPS]

looks consistent with https://www.eteknix.com/adata-sp300-24gb-msata-ssd-review/7/
Now it stalls at 100 BYTES per second.

TrashPanda wrote on 2022-12-04, 17:16:

Wearing out the NAND would take several hundreds of terabytes and would take more than a year of constant 24/7 writes to achieve, even data centers take years to wear out SSDs and will generally replace them before they hit 80% capacity. You should read up on how SSD controllers handle worn out Nand cells.

So unless you are doing hundreds of gigabytes of writes per day on your retro rig I doubt you will ever kill a SSD with it.

You dont need hundreds of gigabytes if your system does lots of small writes all over the shop. Nand pages are big, 100-1000KB big. Its not writes that kill NAND, its Erase cycles. Cheap drives are deeply under 1000 erase cycles endurance. Crap cheap one like that Adata (I found a receipt in my email, $24 in 2014/07) above probably in hundreds range.
As for lots of writes you dont have to look far, couple of years ago Chrome had a weird habit of Caching YT videos despite never reusing that cache, every video watched on YT was slammed to disk cache. Easily gigabytes per day if you watch a lot, and thats just stupid YT.

Open Source AT&T Globalyst/NCR/FIC 486-GAC-2 proprietary Cache Module reproduction

Reply 56 of 56, by pentiumspeed

User metadata
Rank l33t
Rank
l33t

Not true. Typical SSD has less TBW, stress on total writes, to hit the price points, like 100 TBW. Meaning you can wear out the SSD easily if working with SSD heavily. Yes a counter, not years.
You can keep using SSD with exhausted endurance but reliability plummets and easily corrupts the data.

Consider two SSDs.

DC S3500 80GB is TBW of 45 while other end of capacity is 800TB is TBW of 450. An low end SSD.

DC S3610 1.6TB is 10.7 PBW Yes PB. This is considered high end.

This has to do with reserving the unpartitioned space after the partitions size affects the endurance. You can make the SSD last longer if leaving the 200GB unpartitioned on a 1TB for example.

Not all consumer SSD have same endurance. You have to research. Low end and generic SSDs have very small endurance, so consider this a warning beacuse NAND chips are very pricy.
My SSD at work is 512GB Hynix and I consider this a low end SSD and already have wear of 2 already and it was purchased used.

And I can make a SSD have very large endurance by using a 200GB Intel 710 on a PCI SATA card by 3ware on a pentium motherboard and DOS will only use four 2GB partitions.
Intel 710 200GB SSD is about 30 US or 50 CAD each.
Cheers,

Great Northern aka Canada.