@lepidotós The Intel 945 chipset (which can go up to at least a Core 2 Duo) is the last chipset where the ME / AMT does not exist anywhere in any capacity whatsoever. And even then, the Intel P965 and P35 / X38 chipsets do not contain any traces of the ME (or AMT, which is a subset of the ME) either unless the southbridge chip in question is the Digital Office (Q965 / Q35) varient, which is the only version that contains the ME and is then capable (but not to an intrinsic degree) of using the bundled AMT module at all, and even so, there still exist ways to circumvent its ability to phone home.
From what I understand (but have not recently verified; so my memory may be fuzzy on this), the P45 / X48 chipsets (save for one specific ME-free variant) appear to contain the ME by default for offline system operation purposes, but not the AMT module, making the Q45 chipset the only version of that family that carries the potential to be dangerous (which again, can still be easily circumvented if necessary). In contrast, all X58 / Nehalem-based systems may contain both the ME and AMT module by default, but just in an unactivated state when not using the Digital Office-specific chip. Unfortunately, further details are unclear due to Intel choosing to remain extremely tight-lipped about their embedded black box backdoor that nobody wanted.
Further information can be found in the following thread I assembled back in 2020:
For pre-Nehalem systems (Core 2-based CPUs only) that do not use Digital Office chipsets, I'd be much more concerned about Spectre, Meltdown, and the other hardware vulnerabilities that require significant performance penalties in order to patch. Case in point, my primary workstation is currently a Core 2 Quad running on the X38 chipset - with no ME in sight, according to the various methods described in the above thread that can be used to specifically check for its presence.
Anyway, C2D's x86_64 backwards compatibility with modern software (which is almost always dependent on the newest SSE instruction set the CPU revision has) is still a much bigger assurance for most people than the small handful of dedicated developers still maintaining specialty software for PowerPC machines as personal projects in their free time. We won't be around forever - the recent discontinuation of TenFourFox and planned halt of Void Linux PPC, coupled with the slowly reducing number of functional machines in the wild, are small reminders of this inevitable fact - which is why Web 1.1 is such a good idea, because it keeps their online relevance perpetually alive regardless.
-
If you haven't reapplied new paste to your Radeon 9800, I would highly recommend doing so, as well as adding heatsinks to its VRAM chips (more info here). If either the underside of the CPU die or VRAM chips are getting too hot to touch at any point in time and / or its overall video quality is even slightly reduced from what it normally should be, I would even suggest flashing the card's BIOS to the Non-Pro version (the decrease in performance is minor) in order to better preserve the card for more years of service.
ATi clocked almost all of their R300 cards too high for reliability in most cases (opting instead to prioritize a more favorable public image through slightly higher benchmarks), and the vast number of dead or artifacting 9700 and 9800 boards out there compared to other GPUs of the era is solid proof. Therefore, exercising ample caution with these now 20-year-old cards in particular is heavily advised if you wish for longevity.
From what I understand (but have not recently verified; so my memory may be fuzzy on this), the P45 / X48 chipsets (save for one specific ME-free variant) appear to contain the ME by default for offline system operation purposes, but not the AMT module, making the Q45 chipset the only version of that family that carries the potential to be dangerous (which again, can still be easily circumvented if necessary). In contrast, all X58 / Nehalem-based systems may contain both the ME and AMT module by default, but just in an unactivated state when not using the Digital Office-specific chip. Unfortunately, further details are unclear due to Intel choosing to remain extremely tight-lipped about their embedded black box backdoor that nobody wanted.
Further information can be found in the following thread I assembled back in 2020:
A Semi-Comprehensive Guide to The Intel Management Engine and Circumvention Thereof (+ AMD PSP Notes)
For almost two years now, I have avoided all Intel systems with the 965 chipset and newer, instead opting to stay with the 945 chipset and older in order to bypass the security issues surrounding the Intel Management Engine within the unclear 2006 to 2008 general period of initiation. However...
forums.macrumors.com
For pre-Nehalem systems (Core 2-based CPUs only) that do not use Digital Office chipsets, I'd be much more concerned about Spectre, Meltdown, and the other hardware vulnerabilities that require significant performance penalties in order to patch. Case in point, my primary workstation is currently a Core 2 Quad running on the X38 chipset - with no ME in sight, according to the various methods described in the above thread that can be used to specifically check for its presence.
Anyway, C2D's x86_64 backwards compatibility with modern software (which is almost always dependent on the newest SSE instruction set the CPU revision has) is still a much bigger assurance for most people than the small handful of dedicated developers still maintaining specialty software for PowerPC machines as personal projects in their free time. We won't be around forever - the recent discontinuation of TenFourFox and planned halt of Void Linux PPC, coupled with the slowly reducing number of functional machines in the wild, are small reminders of this inevitable fact - which is why Web 1.1 is such a good idea, because it keeps their online relevance perpetually alive regardless.
-
If you haven't reapplied new paste to your Radeon 9800, I would highly recommend doing so, as well as adding heatsinks to its VRAM chips (more info here). If either the underside of the CPU die or VRAM chips are getting too hot to touch at any point in time and / or its overall video quality is even slightly reduced from what it normally should be, I would even suggest flashing the card's BIOS to the Non-Pro version (the decrease in performance is minor) in order to better preserve the card for more years of service.
ATi clocked almost all of their R300 cards too high for reliability in most cases (opting instead to prioritize a more favorable public image through slightly higher benchmarks), and the vast number of dead or artifacting 9700 and 9800 boards out there compared to other GPUs of the era is solid proof. Therefore, exercising ample caution with these now 20-year-old cards in particular is heavily advised if you wish for longevity.
Last edited: