Dietmar

@reboot12 I am just trying your method for the Kingston USB3 stick. This operation needs 4 times longer than my method of to copy the USB stick just as a whole block via Winhex. Do you know, if via your method Winhex makes a check, if the copied data are the same as the original ones? Because there must be a reason for time delay. Still the mystery in the behavior of USB3 sticks in my post above stays. Because there is no error at all after crazy checks, but different structure in the stored data from the USB stick, compared with the structure from a harddisk via Winhex. I use my last acpi.sys from the Ramsey XP SP3 and also the USB3 driver from there (nice for USB3 boot modded driver from @Mov AX, 0xDEAD) on a Gigabyte z690 UD DDR4 board, 12900k with 32 GB ram. Before I noticed real errors in any USB3 driver that I tested, when I copy millions of files one by one from the USB stick to the harddisk. And this errors happen also under Win10, 64 bit Dietmar

Hi, I make a test with two USB3 sticks: Kingston Workspace Datatraveler 64 GB and Sandisk Extreme Pro USB 3.1 256 GB Both sticks show a transferrate of more than 300 MByte/s and until now I am happy with them. Because on both sticks are important files from me, today I make a one by one bit copy via WinHex from the whole USB stick each. Both USB sticks have few bytes left free. And I store this result on a brandnew harddisk WD2003FZEX as*.dat file each. This 2TB Sata harddisk is manufactured in the end of 2023 in Taiwan, solid as much as possible. But now comes my bad surprise: Via Winhex I compare the content of the original USB stick with its *.dat file on this harddisk. Millions of errors are shown. This I have never seen before. When I transfer data between two of those harddisks from 2TB of WD, the error rate is <10^(-14) , means no error(!) in whole 2000 Gigabyte. Just now I am checking, what is going on. I copy the *.dat file back to such another, empty brandnew WD harddisk. and compare the result via Winhex of correctness. Then with with Beyond Compare I check, if there are really errors in the data, or if the crazy USB3 bus copies the space between datas not correct. Just now I think, that even expensive USB sticks are bad as much as possible for to store sensitive data Dietmar EDIT1: As expected, because of the ultrahigh quality of those harddisks, not a single error is shown between the *.dat files and their copy to this brandnew harddisk. EDIT2: The file compare via Beyond Compare 4 shows for 77 different files (all located on the USB stick) the message "stream error, cant compare". So, what does this mean? The error is indeed located on the USB stick, the information cant be read out any longer from that USB stick. Winhex shows during copy no error. In my Winhex versionit is no possibility to check, if a copy has been done successful. Crazy, those 77 files are gone (no, see Edit3). Some big, some small.. EDIT3: Oh crazy, I check by hand some of the files on the USB stick, that cant be read by Beyond Compare 4. They work. All is ok with them. So I run another check with Beyond Compare 2. All files are now shown as identic. What does this mean? The USB stick is not that bad as I thought. Bad is Beyond Compare 4. Nice works Beyond Compare 2 and it is even much faster. And: The files are stored on an USB stick, so that Winhex cant read them out one after the other as on a harddisk. EDIT4: This also means, that you can make with Winhex a correct bit by bit data copy from an whole USB stick. And restore after also correct all the files from such a *.dat file. But the structure in such a *.dat file from an USB stick is different to the structure for the same files(!) via harddisk. Would be interesting if somebody here in the forum can put some light on this behavior of an USB stick.

@K4sum1 Here is the same as Skull2 for server 2003 bit32 https://ufile.io/3a7684oj For XP SP2 bit64 I would use skull2 acpi.sys, because for you it is the one, which matches exact your Bsod. By the way, what shows the first of the 3 acpi.sys for bit64, where I cancel out (dirty hack) the Bsod 0xA5 (0x03, ...) itself Dietmar

@Joaquim Is the Lan activated in Bios? May be, that there is a conflict between the wireless driver and the RJ45 lan Dietmar

@Joaquim I think, you have to install XP SP3 new. And also the harddisk can have Physical errors after loong time Dietmar

@K4sum1 I just see, that the acpi BSOD 03 is in acpi.sys bit64 other handled as in acpi.sys bit 32. Here comes now the crazy hacked XP SP2 bit 64 version, where to 100% THIS BSOD never will happen again Dietmar PS: I find also the reason for this Bsod: C0140008 but I cant locate the exact place, where it happens: ValidateArgTypes or CreateField or ObjTypeSizeOf. This is a really evil Bsod. Maybe idea for hack from Skull is better - 0xA5(0x03, ..., C0140008, ...) DSDT code have operation with unexpected type of arguments, partially solved This BSOD probably means some argument has datatype, allowed only in ACPI 2.0 Patch: - _ValidateArgTypes must always return "OK", even on realy wrong types (mov edi, 0xC0140008=>mov edi, 0x00000000 at head of _ValidateArgTypes) https://ufile.io/sepw9iul and the much more nice solution from Skull https://ufile.io/u42mdknr or this one, only modded for ValidateArgTypes https://ufile.io/463tiqvh

@K4sum1 This kind of Bsod is not possible with this acpi.sys Dietmar

@Joaquim I think, you have to install the Lan driver new. Here it is Dietmar https://ufile.io/nkbuou9w

@K4sum1 Here it is. This was the latest one, that I build from Sources Dietmar https://ufile.io/zkdrmcs7

@K4sum1 I think, that the original acpi.sys has been copied back. This acpi Bsod is now impossible with the last acpi.sys from Ramsey Dietmar

@DaniiX I am not sure, what the reason for the not working keyboard is. May be it is an hidden Bsod, and the compi does not react at all. This can happen, if your harddisk is not correct formatted with MBR, force LBA, ntfs and ntldr. It can also be, that you need a special driver for this keyboard. You can test different USB drivers from Ramsey Dietmar

@user57 You can easy get all those updates via https://legacyupdate.net/ Because this lasts very long, I always use after XP SP3 install the updates first from https://winfuture.de/downloadvorschalt,2136.html and then the updates from Ramsey and the rest via legacyupdate.net Dietmar PS: There is via those updates always the error in the msi installer. This you have to solve by hand via Click Start, click Run, type MSIEXEC /UNREGISTER, and then click OK. Click Start, click Run, type MSIEXEC /REGSERVER, and then click OK.

@DaniiX "So, can I boot it from USB if it's possible for XP?" Yes, I make USB boot possible for XP on any compi Dietmar

@DaniiX Make a try with the nice XP SP3 from Ramsey. Choose the Sata driver from Kai and the acpi.sys from this XP Dietmar

@Mike C I think, that something other goes wrong. Make a try with settings "Write Through" for Ram in Bios. Because "Write Back" halves your memory. And set all the memory timings to the most conservative Dietmar

@Mike C I use the normal IDE driver from XP. The only thing, that I can imagine is, that you have to format your harddisk new. MBR bootsector. I use RMPrepUSB for this, choosing ntldr, NTFS and force LBA Dietmar PS: Are you sure, that your harddisk is still ok? From my about 20 IDE harddisks, only 2(!) are ok after 26 years.

@Mike C Is your harddisk recogniced in Bios? If yes, this configuration should work Dietmar

Hi, I flash this AWARD Bios from 2001, it has also CD-Rom boot for the Shuttle Hot 433. In Bios you have to set boot order CDROM, C. The harddisk has to be an IDE disk (without any adapter), smaller than 137 Gbyte, jumpered as MASTER Dietmar https://ufile.io/d67gsrhn

And here is the same program for XOR with my nice Neuron in Scratch. Oh my, here you can see at once, how bad "higher" program languages are Dietmar

I just make a new XOR program for my Neuron with two thresholds, see post above waaooh, Dietmar

I do an intensive test of those 3 of my Neurons, who make the XOR function without direct calculating XOR. Some thoughts about my idea: No noise at all, no vanishing gradient, always a clear signal, most easy calculation. The output of each Neuron is of type INTEGER. And it accepts any double values as input. Which means, this New Neuron and so this XOR is error tolerant as much as possible Dietmar PS: This program I test under Netbeans 16 and it works. package newneuron; import java.util.Scanner; public class Newneuron { // Inner class for Neuron public static class Neuron { private double lowerThreshold; private double upperThreshold; public Neuron(double lowerThreshold, double upperThreshold) { this.lowerThreshold = lowerThreshold; this.upperThreshold = upperThreshold; } public int activate(double input) { if (input >= lowerThreshold && input <= upperThreshold) { return 1; } else { return 0; } } } public static void main(String[] args) { // Create a Scanner object for user input Scanner scanner = new Scanner(System.in); // Define the neurons with thresholds 0.8 and 1.2 Neuron neuron1 = new Neuron(0.8, 1.2); Neuron neuron2 = new Neuron(0.8, 1.2); Neuron neuron3 = new Neuron(0.8, 1.2); // Prompt the user for input values System.out.print("Bitte gib einen Wert für Neuron1 ein: "); double input1 = scanner.nextDouble(); System.out.print("Bitte gib einen Wert für Neuron2 ein: "); double input2 = scanner.nextDouble(); // Activate the first two neurons with the respective inputs int output1 = neuron1.activate(input1); int output2 = neuron2.activate(input2); // The input to the third neuron is the sum of the outputs of the first two neurons double sumOutputs = output1 + output2; int xorOutput = neuron3.activate(sumOutputs); // Print the results System.out.println("Neuron1 Output: " + output1); System.out.println("Neuron2 Output: " + output2); System.out.println("Neuron3 Output: " + xorOutput); // Close the scanner scanner.close(); } }

I just have a crazy cool idea of a new KI. Take a look at this Neuron. In some kind it is more intelligent than human Neurons in brain. It uses the "All or Nothing" idea for to make a Neuron to fire (first threshold1). BUT: It has a second threshold2, so that it can solve with only such 3 Neurons the XOR problem. And everything with only positiv(!) weights Dietmar PS: My version of Netbeans 16 allows only 1,3 as input for testing for a double number instead of 1.3 package myneuron; import java.util.Scanner; public class Myneuron { private double threshold1; private double threshold2; public Myneuron() { threshold1 = 1.0; threshold2 = 5.0; } public int activate(double input) { if (input < threshold1) { return 0; } else if (input < threshold2) { return 1; } else { return 0; } } public static void main(String[] args) { Scanner scanner = new Scanner(System.in); Myneuron neuron = new Myneuron(); System.out.println("Test meines speziellen Neurons:"); System.out.print("Gib eine Kommazahl als Input für das Neuron ein: "); double input = scanner.nextDouble(); int output = neuron.activate(input); System.out.println("Antwort des Neurons: " + output); scanner.close(); } }

@Zdzichu Search Internet for Ramsey XP Dietmar

@reboot12 I take a look at it tomorrow, thanks for help, now I go to bed Dietmar

Hi, I just test your bios image.bin. No changes compared with original Bios. Amisetupwriter tells Write Protected. And CSM is shown as disabled and Write Protected as enabled. No menu for CSM in Bios. I set before the jumper for CMOS. ru.efi tells the same. Dietmar