pointertovoid

The Agp aperture is just a part of the Cpu's address space (and I believe of the mainboard physical Ram, at least when you don't have Fast Write settings for Agp on both the Gpu and the chipset). So this won't limit the amount of Ram the Gpu can and will use on its card. Typically, the Gpu stores textures locally before playing and gets texels through the Agp in real time.

Hello everybody! UniAta driver is a unified driver for many disk hosts: Pata/Sata/Scsi, Intel/nVidia/Via/etc, Nt4/2k/Xp/2k3/V/etc... Which makes it very useful. http://alter.org.ua/en/soft/win/uni_ata/uni_ata.php BUT its Inf file doesn't mention the ich10r disk host of my computer. When I give the UniAta driver to the W2k installer as an F6 floppy, the installer takes the files without any comment, but at the end I get the usual Intel driver supplied by the W2k installation Cd. So: - Does UniAta work on ich10r? - Then, how can I convince W2k's installer to use UniAta? A modification in the Inf file I suppose? Thanks!

I would also begin with PtEdit32 (just for not having used TESTDISK up to now). It can display the partition and volume information without touching it. It allows to save this information before your try any repair, like forcing the few bits who tell "this is Fat32". And it allows to revert the volume information you saved before, in case the changes you made didn't improve. So this is a rather harmless option and I would consider it before other methods like a backup, whose effects are less predictable.

Just from Win95guy's name, you may infer he's aware of extended support and out-of-support OS... Personally, I don't rely on Microsoft for security software, as for instance their firewall for Xp is badly rated. So I use third-party security software, and these won't expire in July 2010. Maybe sooner, maybe later. There are still a bunch of them, and some of the very best ones do run on W2k. Win95guy, I appreciate your inputs for my W95. Regards!

I've already read the same complaint about Flash Player on another forum. Seems to relate to full screen rather than to the number of pixels. And Flash Player just sucks anyway. So no, it's not related to a specific hardware. And any decent piece of software should be able to display a video on your computer, which is able of running 3D games.

Yes, there are some reserved places on the disk that describe the volumes. The beginning of the disk tells where the volumes are, how big, and which one to boot. Each volume then tells its own type and may contain something bootable. http://en.wikipedia.org/wiki/Disk_partitioning And PtEdit32 allows you to read and write in hex exactly this kind of information, and gives some interpretation. So if only the information "this is Ntfs" is missing (...would be a piece of luck) you could try to restore it.

Weird. I tend to believe what GParted tells. You could try PtEdit32. It gives more low-level information than GParted and allows you to set the partition information without touching your documents. ftp://ftp.symantec.com/public/english_us_...es/PTEDIT32.zip on W2k it requires admin session, on W98 you must add Pqvxd.vxd taken from vfd_folder.exe at Symantec. From what you describe, I guess the Promise driver is installed. Anyway, the Promise adapter works without an added driver as well. And this wouldn't explain the slave disk on the mobo's host. This reminds me - but it isn't the right explanation - of the special sector Promise takes on the disks to keep Raid information. When HdTune and colleagues write on this sector, you lose the Raid. Rebuild it identically, you get your data back. Next time, "low-level-format" the disks attached to the Promise host, and then this sector is protected. Or use a Sil0680a host, it's better than the Promise in any aspect.

And does this latest Bios bring a new Ahci version? It's written when the Ahci part gets control. For the ep45-ud3r, only Ahci 1.07 and 1.20E are available, the latter and faster one is embedded in Bios F11 only.

You will have to separate the printed circuit anyway. So if there is no visible damage, why shouldn't you try to clean the printed circuit and look if it works again, before replacing it? Having powered it when wet of course didn't help... and if water entered the platter compartment, I would say it's over, as the platters are dirty now. But anyway, these are the directing lines: Water deposits dirt (salts) that create electric path, conductive enough to disturb electronic circuits. These deposits must be removed. First, rinse with plenty of water. You may begin with tap water, but you must end with deionized water, as is sold for lead batteries. Clean the container before putting deionized water, and try not putting you fingers in it. Of course, electronic parts shouldn't be left for hours in water. Then, rinse with pure alcohol, or with an alcohol mixture used to clean windows. The aim is the same: no dirt left after evaporation. At the end, dry up with a hair dryer. ----- You can find a printed circuit at eBay, by buying several broken disks of this model (called there "untested" or "sold as broken" and the same...) until one has a broken platter and a sound printed circuit. You can even find the printed circuit alone there. eBay Dot Com and Dot De only, others are too small (among other reasons...) A more secure method, as you probably won't use the damaged disk after you retrieve your data, is to buy a working disk of this same model and take temporarily its printed circuit to put on the old platter. After retrieving your data, put the circuit back on the working platter, as well as the data. Or sell the working disk after testing it. ----- A few companies specialize in retrieving your data. Not really cheap, but less expensive than one imagines.

One efficient method is to put the laptop on raisers, so that air moves easily under it. Some are even sold for real money. You can also find flat blowers designed specifically to put laptops over. However, a computer should be able to work normally... Check that all fans rotate normally, air ducts+fans+exchangers are free of dust, and voltages and frequencies are normal.

I too confirm the existence of an upgrade Cd from fe to se. I got two Cd, the 98fe and the upgrade, with a new computer. This upgrade Cd doesn't install alone BUT can repair alone an existing 98se installation, because it contains all files. No other key is required. Its organization is nearly identical to a standalone 98fe or 98se install Cd. W98fe and 98se use the same keys (and all languages as well, but oem differ from boxes), so for practical uses (which is not the thread's original question) one can take a full 98se and give it a 98fe key. This limits the usefulness of the upgrade Cd. I also heard from an Internet Provider's hotline that one-shot 98se is better for installing modems. By the way, W95b also uses the same keys as W95a, in all languages. I couldn't determine if using a key supplied with 98fe to install a 98se is legal. No information at Microsoft, even then. The Oem sticker just tells "licence for 98". Same for 95. Disks +licences sell at eBay for some 5 euros including shipment. In all European countries where Justice had to decide, software can be resold used, even without hardware, even in Oem versions. In Germany, Justice logically deduced that technical means needed to install a Compaq Windows elsewhere are legal as well. After half a dozen convergent decisions, software makers avoid asking to Justice in more countries here, and hope customers will go on believing that the so-called licence agreement is law.

And it worked. I just used Gigabyte's regular QFlash to update the mobo's Bios from F4 to F11. This brought the embedded Ahci part by Intel from 1.07 to 1.20E, which now takes 3s instead of 17s. So I didn't have to modify a Bios, which I was reluctant to do. Two other parts of the computer start have gotten each 3s slower: before the Bios gets the screen, and during Xp bootup. But the net sum is clearly positive, and at 4000MHz, I get a working Xp 25s after switching the power. Still a little bit disappointing. On my PIII, the Bios passes control to Windows boot process some 8s after I switch the power on. Are there P45 mobos with a faster Bios than my Ga-ep45-ud3r?

Found some threads on diverse forums that observe the same effect, and their analysis and results are quite clear: http://www.insanelymac.com/forum/lofiversi...hp/t168690.html http://62.109.81.232/cgi-bin/sbb/sbb.cgi?&...1&show=9789 It was really Intel's embedded Ahci that slows down Gigabyte's Bios, and Gigabyte took some time to include Intel's faster v1.20E Ahci. So for my Ud3r, official Gigabyte F11 Bios should be the answer. And of course, I'll tell you the result.

Meanwhile I've downloaded Bios F4 F5 F9 F11 from Gigabyte, detailed them by Cbrom v2.20 and observed the embedded Ahci subroutine provided by Intel. F4 F5 F9 include Ahci v1.07, only the F11 (from 31st of August, 2009!) has Ahci v1.20... Which means that I can hope a classical upgrade to F11 will bring a faster Ahci part of Bios start.

Err... Humm... Cough, cough. Well, after installing Xp in dual-boot after W2k, I just noticed inadvertently that Ntldr and Ntdetect had changed their version, but as Xp needs its own files, I couldn't revert to W2k's ones, you see. And then, when measuring boot times, 5.1 file versions were in place for W2k as well, which makes the comparison unfair, as I would revert to 5.0 if using just W2k, you know. 2k3 files seem very close to Xp on my mobos.

I use Ahci with the X25-E because this drops boot time from 35s to 9s. The reason is very unclear to me; once Windows runs, I get full throughput and response time in P-Ata compatibility mode; only Ncq performance differs. I don't believe Ncq alone does this improvement in boot time. Rather, the P-Ata compatibility mode must run at Pio4 or Mdma2 at the beginning of Windows' boot. The long Ahci part of Bios setup doesn't depend on the connected drives, from my experiments. It is supplied by Intel, but from answers I got from a German forum, versions 1.20 and newer are fast (<1s) and Gigabyte hasn't included them in the Mobo's Bios. I still have to process the heavy information there, solutions would imply replacing the Intel part of the Mobo's Bios by a less old version.

Right, o2003 can get its updates slipstreamed properly. That's an advantage. o97 is touchy with its updates, can't officially be slipstreamed and needs updates to be applied again if you change optional components. You may even need to de-install some updates before. Still not a decisive advantage to my eyes, knowing the price of o2003, but it is an advantage indeed.

In an attempt to understand the difference in boot times between W2k (17s) and Xp (8s) on this same hardware (X25-E, ich10r, P45, E8600) I've tried to use BlackWingCat's modified Ahci driver on Xp as well. Result: Xp still boots in 8s. A strong clue that the original and modified Ahci drivers are equally fast, and Xp boots inherently faster than W2k does - at least on this hardware with Ncq and 2GB Ram. I had already seen Xp boot quickly on a 7k160 +ich2 +i815ep +PIIIs +512MB. I got the impression that Xp started faster than W2k does, but the comparison wasn't fair then (old installation vs new, naked installation).

Ciao Jaclaz! Hum, I see what you mean... And unfortunately, I did keep notice of all adresses for downloads, but BlackWingCat's site has moved meanwhile. This page looks like the one I used: http://blog.livedoor.jp/blackwingcat/archives/813816.html even better, it already was called 813816 then. In this page, ダウンロードlinks to http://files.me.com/theblackcat/y53fv3 which proposes a file of nearly the size I downloaded, corresponding to Intel's v7.6.0.1011. Since 2009年04月17日 it may have changed on 8th Sep, 2009: Fix add Intel® ICH9M-E/M SATA AHCI Controller If you download it (20MB, I didn't do it again) it shall look like a usual Intel driver, except that for instance the contained Winall\Driver\IaStor.sys has an invalid signature (logically enough) and an added CustomizerCopyright. I checked (for a previous version) that it calls only entries available from W2k, as opposed to what Intel's original does. Well done! Maybe BWC passes by and tells us more... Did you notice? Internet addresses will be allowed to contain non-Latin letters. This will certainly make life much easier for Japanese people.

Hello everybody! I have a Ga-ep45-ud3r mobo with ich10r host and Bios version F4. If Sata disk access is set in Bios to P-Ata compatibility mode, then W2k-Xp bootup is slow (35s) for still unclear reasons. And if Sata disk access is set in Bios to Ahci, then OS bootup gets very fast (OS disk is an X25-E, W2k boots in 15s and Xp in 9s, including nVidia video driver). But then, in Ahci mode, the Bios gets very slow. The video card displays its data for a few seconds, and then the Bios takes 17s before giving control the the OS boot sequence, which is far too long. From what's displayed, it seems that Gigabyte's Bios calls a routine supplied by Microsoft, as soon as the ich10r is set to Ahci or Raid, and this routine takes about 15s. Right? By the way, this mobo has also a JMicron 363 P-Ata+Sata host on board, which also slows down the Bios pretty much the same way when in Ahci mode. So: - Can I improve this time through some settings? - Does your mobo's Bios run faster than 17s in Ahci mode? What's your model then? - Does an Intel mobo, with an Intel Bios, run faster than 17s in Ahci mode? Observed figures? - Or do nVidia chipsets manage it better? Thanks!

My experience with Hitachi drives is excellent, as well as with Excelstor (technology licensed from Hitachi) - though I don't have an experience of 20,000 dead drives... And for Laptops, Hitachi have (among the) fastest drives. Prefer a 7200rpm, it makes a huge difference. And as big a capacity per platter as possible. Find as many P-Ata drives as you want at eBay. Did you test your drive with the appropriate software? Drive Fitness Test is there: http://www.hitachigst.com/hdd/support/download.htm just in case a trojan has overwritten your hdd's boot sector. ----- Malware creating hardware failure: maybe... Because many firmwares can be flashed from within Windows, which I don't see as an improvement because of this security concern. Also, settings of chips can be changed from within Windows, as is done for Cpu clock for instance. Changing the temperature alarm as well as the frequency or the supply voltages could be harmful. Zap the Ram for instance. As far as I know, any adjustment done by the Bios can be done from within Windows as well; it may need an account with administrator rights, or may need to install some driver or service. Such malware would need to be sophisticated and maybe specialized on some hardware, but sophisticated malware already exists.

One possible cause would be if your current power supply is a pre-Atx one (corresponds to P1 vs PII era). Atx introduced improvements to allow wake-on-lane; especially, they supply a weak, separate, permanent +5V to all devices that are allowed to wakeup the PC: mouse, keyboard, Lan cards, modems. An Atx power supply lets some peripherals react when you switch the current on but the Bios starts only with the "Atx power" button. Well, normally. And an Atx power supply disconnected from the mobo doesn't rotate its fan until you short two special pins at the output. I didn't check if the D-Link 520Tx PCI has the (common) capability.

Oops, that's a bad joke from Microsoft! I hoped to answer with the download address at MS, and apparently the Sp6a has disappeared from their site! Its name is SP6I386.EXE, and the French address was http://www.microsoft.com/downloads/details...;DisplayLang=fr many users complain in all languages. Googling did give some answers outside MS for English version, but I couldn't find any Dutch one. Hope someone has stored it. By the way, MS will stop maintaining W2k in July 2010, so it's time to store all updates on your computer before they know the same fate. Same for Office 2003 (in October 2009?).

Does your W2k already have the Sp4? It adds a lot to its Usb capability. ---------------- EDIT : so sorry! The issue was already solved and I added an irrelevant message. Mea maxima culpa!

And I could install W2k on an X25-E connected to an ich10r, with Bios set to Sata-Ahci, using BlackWingCat's driver on an F6 floppy. The floppy had the contains of BWC's Winall\Driver folder, without any change, and it just ran smoothly. Well done, BWC! My W2k installation Cd is heavily HfSlipped to the December 2007 status, but includes no other driver than those brought by Sp4 - which means, this Cd itself ignores everything about ich10r and Ahci. Disk throughput is excellent with BWC's W2k driver, as well as NCQ: IOMeter tells random 4k reads jump from 6000 IO/s at Q=1 to 40,000 IO/s at Q=32 just as Intel advertises; Atto tells similar improvements; and the Ncq doesn't work without Ahci mode and adequate driver. This W2k boots much faster (17s on X25-E and E8600) than if installed and booted in P-Ata compatibility mode, but not as fast as Xp with Intel's F6 (8s). It's perfectly possible that Xp boots inherently faster than W2k does, especially as the prefetch mechanism combines well with Ncq. Still investigating.