dencorso

7:00 PM Pacific it'll be in 15 minutes... OK, let's do it! Now, for a starter, give me please two screenshots: on from HD1 and one from HD3, showing the errors. Let the highlight on the MBR (where is by default as it starts) for these 2 screenshots. We'll be using solely RPM 2.43, OK? The error you mentioned is "MBR/EMBR partition overlaps an existing partition. Partition not imported", right? Take good care not to modify anything in HD1, for now.

Well, in GB you cannot do it, but in kB it's possible to do it with RPM alone. Listen, piikea, do you have perchance about 2 continuous hours to dedicate to it, we might solve this problem for good. I mean, I can guide you step-by-step throught the process, provided that: 1) You're available to do it today 2) You're willing to post many screnshots, one after each step 3) You undetake not to give up before we finish 4) You arm yourself with lots of patience What do you say?

Happy Birthday to both of you, Punto and Trip!

There's a reason why the page pointed to by BlouBul is entitled: "Double-Clicked Documents Slow to Open In Word"... I've been playing around with DDESpy for the last few minutes, and I'm convinced DDE is only invoked when one opens a document by double-clicking on it. If, conversely, one opens Word, then goes through the "File" menu, "Open" routine, DDE is not involved at all. So let's find out whether DDE is the reason of your problem or not, in a very scientific way. You just reported 7 timings of opening the file in Vista, by double clicking on it. So, please do it again 7 times, but now, with Word already open, by going through the "File" menu, "Open" submenu, selecting the file in the file browse box and starting the stopwatch on clicking the "Open" button in the browse box. Then close just the file, not Word, and repeat. I bet it'll be much faster. And then we'll know the problem is with DDE. Or we'll disprove it.

I imagine you don't currently use neither config.sys nor autoexc.bat, is that right? I don't wanna know that you didn't personally create any. I want to know whether those two files do exist in the root folder of your boot partition. If they do, please zip and attach them or post them if they're just a couple of lines each. From there we may solve it to your satisfaction, in a very simple way (well, sort of...), using menus, if all goes well.

So we have 3:26 min (median) for Word on Vista and 1:54 min (median) for Word on 98SE!!! Wow! So, it turned out that 2:21 min (the 1st try) was in fact biased to higher times, not lower ones! And we don't need to work with gut feelings anymore: 98SE is almost twice as fast! Now, that's a quite definite "faster on the slower machine", so the processor upgrade must be ruled out. And your interesting result with Works on Vista suggests Vista itself is OK, too. So BlouBul is right on the mark. We now must look for "something else", and the place to look is in and around Office itself. The link provided by BlouBul is a great find, and IMO outlines a troubleshooting path worthy of being pursued. allen2's suggestion about repagination and any other idea that actually involves disabling some part of Word's functionality must be left to be tried last. This is my 2¢ right now.

Possible answer #1 On computing. More specifically, on computing as related to MS software. Possible answer #2 Good advice, earnestly given. And yes, it does matter. BTW: Welcome to MSFN!

You're failing to see my point, BlouBul. If it were slower in the P233 machine, it might be processor dependent. If it's really faster in the P233 machine, it's probably something that's causing the delay in Vista, despite the higher processing power. But if it takes the *same* time, it may be processor and memory independent, but intrinsic to the way Word processes the file (if, and only if, both versions of word employed do it the same way, that is)... That's what I'm after!

Well, first of all, my gut feeling is the same as yours: it's not the processor. Now, it's sure not a scientific benchmarking survey, inasmuch as we're comparing things done in different OSes and most probably different versions of Office. But repeating mesurements a significant number of times and taking averages or, much better, medians (because they are robust), is just scientific enough, and, although somewhat tedious, not so hard to do, IMHO. And it'd bolster a lot our confidence in the significance of the result. I do quick and dirty tests all the time, but since this is a quite perplexing issue, I feel it's plenty warranted to do it as scientifically as the PITA factor allows. Of course, that's just my 2¢.

Sorry, but I remain in doubt. 2:21 min is not so far from the shortest time you obtained in the main machine, JorgeA. While there the values seem to cluster around 3:00 min, this might have been an incredibly lucky single try. To believe it's really different, I guess some more tedious work is needed. We only have facts when they are reasonably validated... So, I suggest you time the opening of the big document 7 times in the Pentium 233, and further 7 times in the main machine. And, then, let's compare the *median* value of each set. If it turns out that the median value of the opnening times in the Pentium 233 is really smaller than in the main machine, then I'd say BlouBul's conclusion is warranted.

Does this and this help?

You're right, of course! I had the E4500 in mind as I wrote that, as a bare minimum (larger cache, all the rest equal). But I must point out that there may be some frequency restrictions for JorgeA's board... This info is from HP's page on the Benicia: Motherboard supports the following processor upgrades: •Intel Core 2 Quad (Y) Q9xxxx •Intel Core 2 Duo (W) E8xxx •Intel Core 2 Quad (K) up to Q6600 •Core 2 Duo E6x00 (C) up to E6700 •Core 2 Duo E4x00 (C) up to E4400 But if it accepts the E6700, I don't see why it shouldn't accept anything above the E4400! It must be a typo.

1) It's not in the registry. It's on section [386Enh] of SYSTEM.INI! Mine reads: PagingFile=J:\PAGEFILE.SYS PagingDrive=J: MaxPagingFileSize=1835008 MinPagingFileSize=1572864 2) There are *many* ways to do it. 3) Why not? If they're FAT (12, 16 or 32) DOS ought to recognize them all right. BTW, do read this.

Well, I've been thinking and, if I were you, I'd partition your disk with a single primary active 30 GB FAT-32 LBA partition, followed by an extended partition containing 3 logical FAT-32 LBA partitions of about 300 GB each. After formatting, I'd try to scandisk them with the windows version of scandisk, and if all goes well, start using the disk. Only in case the windows version of scandisk complains "there's not enough memory to run", then I'd repartition to smaller partitons, but I do believe the scheme I just proposed can work OK. The primary active 30 GB FAT-32 LBA partition would be left there for that possible future OS installation you mentioned way back in this thread. Alternatively, a single primary active 250 GB FAT-32 LBA partition, followed by an extended partition containing 3 logical FAT-32 LBA partitions of about 250 GB each is also viable. But do not use 4 primary partitions, each of about 250 GB, because 4 primary partitions is allowed but very non-standard, at least for MS OSes.

A 60% CPU usage for assembling a text document is a fairily high usage. So, probably, the best way to improve your loading times would be a processor upgrade/ Considering your motherboard, a drop in replacement at a reasonable price would be a Core 2 Duo E4400 (for about $130). More powerful processors would cost much more, and entail the need for higher speed RAM, making it too expensive to be worth it, IMO. But if you think it worth it giving a shot at it, CoffeeFiend and PuntoMX are much more knowledgeable than myself about Intel processors, and probably can give you better advice. I use 0.5 GiB on Win XP and 1.5 GiB in Win 98SE (I have just 3 GiB RAM, and my XP is much happier with 2.5 than with 1.5 GiB available). I wrote it in more detail, with 98SE in mind, here .

Thanks! 1850% faster sequencial read and a lot more in the random read! Wow! So it's valdated, all right! Hence, it's official: storage device speed doesn't influence your issue at all! And that answers your original question: No. An SSD won't help at all. I was suspecting you'd feel that way. So would I, were I in your position, too. You might now invoke the CPU Meter, send it to the tray, open word with no document, then use the file/open menu to select your big word document, and monitor the total CPU usage during the 3 min it takes to load. Or you might not send it to the tray, find the line for WINWORD.EXE in the "Windows Task Manager" Process view, and follow the actual CPU usage of Word, during the loading. That should tell us how much of a CPU drain loading this big document is, in fact. And doing it both ways should tell us whether it's Word or other processes that are doing most of the processing. As for trying to open your big document in the Win 98SE machine, that should be an interesting experiment, too. It's sure worth trying.

OK. Here's why I said I don't think it's necessary to do a clean install, to test whether Norton has a hand on it. Symantec provides anyone for free the Norton Removal Tool, which does a pretty thorough job. However it removes all Norton products from the machine at once. So the downside it to have to reinstall all Norton products one wants back, after the testing, all over again. For one or two products only (and that's why I asked how many), that's feasible, and much less work that a reinstall from scratch. Now, to make clear what I think about Symantec products, it's like this: I *love* Norton Ghost 2003; I like a pretty lean install of Norton System Works 2003 up to 2005. I think all later products from this line are very much bloated... and I abhor the Norton Antivirus, which I've always seen as a big system hog. Then again, I like a tweaked installation of AVG (the paid-for AntiVirus, not the Internet Security), but it seems I'm almost alone in liking it, these days...

@PuntoMX: Sorry if I came out harsh! No offense was intended. Please do keep around, things are getting "curiouser and curiouser" just at this point.

Let's validate the test first: download Crystal DiskMark and benchmark both your HDD and then the ramdisk (screenshots welcome! ). The difference must be huge. If so, then really the bottle-neck is not loading the file, but processing it. It just may be that word really needs all that time to assemble the document from the way it's stored in the file. PuntoMX abominates Symantec Norton products, but he may have a point there. However I see no need for a reinstall, at this point. Tell me, please, how many Symantec products, Norton or Corporate, do you have in your setup, and which products are they.

RLoew's hypothesis is easier to test than mine. So, get Japheth's HIMEMX.EXE and check whether the machine boots OK and starts the GUI OK. If so, then do a scandisk on the ramdisk, using the Windows Explorer Properties for the ramdisk, and direct it to do a surface test. If RLoew is right, you'll be able to get to the GUI, but the surface test will crash the machine.

If all other hypotheses can be safely discounted, then let's go back to the faster storage device idea. I'd say create a Restore Point, then install the Free Version of VSuite Ramdisk. It's the faster one available. Set it to 512 GiB. Then copy your test file to it and do a test. Compare the full load times, an let's see what happens. It doesn't hurt to try. And if it actually solves your problem you might even consider upgrading to the paid version, which is even faster, because it has a Direct I/O mode, although it may not be worth, perhaps, paying for it. I bet the free version will be enough, already. But let's see. Edit: I overlooked the fact that you use a 64-bit OS. This means you have to use the paid for version of VSuite Ramdisk. However it's 15-day free trial, so you can test it all right. Now, if you want a free Ramdisk, there is the Dataram or the Gavotte. And both are good enough, too. However, the former is easier to set up. Then again, the VSuite is twice as fast, at least, because it can use Direct I/O.

Are you using an English or a Polish Win 98SE? In any case, I'd say disable fully both ACPI and APM. At least for testing.

You mean "live", of course, don't you? @HardDriv'n: Try upgrading the Windows Explorer to this one (version 4.72.3612.1710, US-English).

The whole thread at Boot-Land I pointed you to is dedicated to just one of the many ramdisks available: Gavotte's... Ramdisks both as a concept and as actual software are alive and kicking. Here's still more to chew on: Ramdisk Benchmarks. That said, I think CoffeeFiend is right on-the-mark: probably your problem goes well beyond using a fast storage device. The delay you reported seems to indicate a lot of processing going on. You should investigate it further. However, since setting a ramdisk can be done for free, it might help comparing what happens on opening some of these files from your current HDD and from a ramdrive, which is way faster, to establish whether the speed of the storage device actually does bear on it (and if so, how much) or not.

If money is not a problem, you should go for a Fusion ioDrive. If it's somewhat of a problem, a RAID array of 6 class 10 SDHC cards will do nice. But usually both are too expensive as solutions. For normal mortals, A 4 GiB Gigabyte i-Ram is the best solution money can buy. But you'd have to load it the files at the start of a working day and backup them to safer media at the end of the day, just to keep on the safe side. All going well, the files keep there all right, so you just need to backup at the end of the day, each day, to keep safe, but the files are still in the i-RAM when you start next day, so there's no need to reload them. I'm a quite satisfied user of i-Ram, although I use it just for the pagefile, most of the time. Read much more about it and details about the exotic media I've mentioned above in this thread elsewhere (posts #38-49... it was sort of a thread hijack, as you can see). Or you could go the ramdisk way: a 256 or even 512 MiB ramdisk could be easily created from your existing RAM, would impact little on the performance of your Vista system and is surely faster than any of the above mentioned hardware, except, perhaps, the Fusion ioDrive.