|
|||||
Kickass Phone Rates from 3U
|
|||||
| PLACES TO GO: |
BIOS Settings Manual Description and Translation by Wayne "Hat Monster" Hardman Publisher's Note: This was originally published on ArsTechnica and it is such a great reference that I use it all the time. I figured I might as well put it up here for others to find and use. I've edited the text a bit, the format, and added in quick links for people to find notes on the sections of the BIOS that you're looking up. All credits for the content go to Wayne Hardman. Quick Links to Content The BIOS is that dark, cavernous pathway to your computer's very soul that both beckons and repels the would-be performance seeker. The BIOS can open up major performance pleasure for you, or it can hide that one little pesky setting that's causing your daily bluescreen, and you just don't know why. If you royally jack your BIOS, you can toast your system to the point which you may need a system reinstall or some new hardware. Don't get me wrong--it's not as though it's all that easy to do major damage to your system via the BIOS. It's just a good idea to know what you're playing with and not change things around willy-nilly. Most PC users don't even know what a BIOS is, "where" it is, "how" you "see" it, and what you can do with it. That's hardly surprising, because most users purchase OEM machines, and OEMs aren't too excited about users tinkering with their BIOSes. That's one of the reasons why the BIOS on any given machine may or may not have most of the settings I'm about to talk about. That's also why we need to be clear: we cannot take responsibility for your system should you choose to attempt any modifications based on the material herein. While we have done our best to make sure that this information is accurate and applicable, we cannot be sure how the BIOS on your system is actually set up or how it specifically implements some of the options. You can completely hose your OS install if things do go wrong. But hey, you're reading this, so you already know that performance seeking isn't for the lazy nor the timid. A word on what a BIOS is. Back in the dark days of AT and 286es, IBM were facing problems of their architecture already. Systems were implemented in different ways, and there had to be a way to make the operating system know what's going on. Back then, we didn't have drivers and HALs and needed something actually in hardware--the BIOS. Meaning Basic Input/Output system, it was exactly that, a standard way of getting at the hardware. Long before the term "API" was born, BIOS was a basic API. You could call BIOS ROM routines to handle video, PC speaker beeps, the keyboard, and so forth. Operating systems interfaced with the system like this for a long time. Then, in about the 386 days, another thing happened with BIOS. It had long been used to enter and/or correct CHS data for hard drives, but it now became used to set up chipset registers. RAM timings could be altered, and other innards of the chipset could be tweaked since there wasn't a real way to find out what timings the memory wanted. Remember, in those days, people didn't build their own systems, so it was left to the experts who knew what each cryptic name meant. This brings us to today. The chipset is configured by the BIOS on boot. PCI configuration registers are hosted by the chipset, and these registers are altered by BIOS according to what the motherboard manufacturer has specified, together with what the visible BIOS settings dictate. Everything in BIOS alters something to do with the chipset's behavior. BIOS data must be RAM--you can't alter ROM on demand like that--so a CMOS (complementary metal oxide semiconductor) RAM chip was added. This CMOS RAM was battery-backed and stored the settings that BIOS would use. The battery also kept the real time clock (RTC) going when AC power was off. Today, there is no CMOS RAM chip, but out of habit, we still treat it as if there were one. The CMOS today is actually a small amount of RAM in the chipset that remains powered by the battery when system power is off. By booting into your BIOS, you get a nice, semi-English interface with which you can manipulate those binary codes in the registers. Some settings you can't alter, some you can. Some aren't even present. Those you can typically tweak on an enthusiast motherboard (non-OEM) are covered here. Those you can't are either really, really nasty voodoo hidden by your mobo manufacturer, or just not relevant. This guide is divided into three major sections, and the subsections are ordered alphabetically:
The RAM options are some of the most difficult to understand, but potentially the greatest performance boosts the BIOS can give you. The sharp amongst you will read "greatest performance boosts" as "easiest ways to lose stability" and you'd be right too, in some cases. If you're intending to max out all these settings, you're going to need the RAM and motherboard of the Gods. That's not to say you can't compromise and tweak up some settings but leave others more relaxed. That's what the art of tweaking is all about: finding that sweet spot. Keep in mind that the chipset is, after all, essentially a go-between between the CPU and the system RAM (as well as PCI, AGP and the like). It typically holds true that higher overall RAM bandwidth is indicative of a better performing, more advanced (read: recent) chipset. It is very much possible to tweak out your chipset so much that it will outperform the 'next one up'. For example, a heavily tuned PC133 system can run a very tight race and even emerge triumphant over un-optimized DDR200 systems. My own KT7 with the KT133 chipset will handily beat a 200MHz AMD760 system with DDR RAM, even thought I'm only using SDR RAM. You can do the same. Remember that Quake3, 3DMark2001, and many other 3D graphical glories scale with RAM bandwidth almost as much as they do with video card blastage. Take a look at some of these with your bootable memtest86 CD (c'mon, nobody uses floppies) in one hand and Prime95 in the other.... Options: Enable, Disable This should always be disabled for modern systems. In the old days, ISA cards used the 15th megabyte for their own purposes. Modern ISA cards don't need this (EISA or PNP-ISA) and neither do PCI cards with one major exception -- Sound Blaster PCI128 and Sound Blaster Live cards like this to be enabled. It essentially removes 1MB of your RAM, so consider replacing the sound card instead. The SB16 emulation of those two cards is the business here -- don't install it and this shouldn't be a problem. Nevertheless, it's another suspect for crashes associated with the Sound Blaster.To replace SB16 emulation, try using VDMSound, which gives emulated sound for DOS programs in NT4, 2000, and XP, and doesn't need any special drivers. Options: Enable, Disable This is identical to Spread Spectrum's "Smart" function. It turns off slots that aren't populated. You don't lose anything by enabling it, and it may slightly reduce power usage. Options: Disabled, 2-Way, 4-Way The chipset can interleave requests to different DRAM banks and gain a massive performance boost from it (relatively speaking). Without interleave, a typical session would be:
All four banks can be accessed in turn, without having to refresh the RAM and wait the killer CAS. This is four-way interleave. Two-way interleave is only two banks at a time. The only time you will be unable to set 4-way is if you're using a very small amount of RAM. Every DIMM has at least two banks, but 64MB and smaller DIMMs have only two. All larger DIMMs have four. However, if you use two two-bank DIMMs, you have a total of four banks and can use 4-way. Always set this setting, then, to 4-way and don't expect any problems from it. Bank interleaving was not covered in the RAM guide, so I've gone into a bit more detail here. Options: ECC, Non-ECC ECC has to be supported by both the chipset and the RAM. If you've shelled out the cash for ECC RAM, you had better use it, right? There is a small performance loss associated with using ECC in exchange for the error correction. You'll know yourself whether you need to have this set to "ECC." If you don't know then it should be set to Non-ECC. Options: Auto, No Delay, 0.5ns, 1.0ns, 1.5ns Marginally different from the one that only has Enabled or Disabled, this tunes for high RAM loads. One single-sided DIMM isn't a high load. Two double-sided 256MB DIMMs are. For lower RAM loads, lower the latch and have it at no delay. With more DIMMs, you need more of a delay or the chipset may find itself unable to properly latch onto a bank. A higher latch reduces performance. If you have crashes after upgrading your RAM, this is the very first setting you should look at. Options: 2, 2.5, 3 This is the famous CAS latency. It's part of the wait between the chipset requesting data and the DRAMs getting ready to send it. A shorter delay is better; 2 is less of a delay than 3. However, your RAM needs to be able to handle the increased rate and may not be able to do this if the FSB is overclocked or the RAM is of a lower specification. Increasing the CAS latency will therefore allow you to overclock the RAM further (if the chipset and CPU will let you). 2.5 is only available with DDR. Options: 5/6, 6/8 Like the CAS latency, a lower setting is faster. Like the CAS latency, it's more stressful on the RAM when it's lower. Hannibal's excellent RAM guide will take you through this and all other RAM-related jingo. A low Row Active Time will force the data out of the RAM sooner, but it may not leave the row open long enough for transactions to complete, in which case you either need to get faster RAM or increase the setting up to 6/8. Options: Disabled, 0ms, 0.5ms Specifies the time to wait between interleaved transactions. Disabled and 0ms are the same thing and are the fastest. You know the drill with these RAM settings. Give it that 0.5ms if you're using three or four sticks of high capacity RAM just to keep the chipset playing nicely with the high load. Options: 2, 2.5, 3 The third part of the x-y-z notation used in SDRAM, the other two being CAS and RAS to CAS. Like its brethren, it's better lower but also more stressful lower. See the pattern? 2.5 is only available with DDR. Options: 2, 3 When RAS is asserted, there must be a small wait before the CAS can be pulled. This setting controls length of the wait. Like CAS latency, it's a delay before you get your data, so while your system is faster at a lower setting, it's also more stressful at that setting. Your RAM may handle it, or it may not. Options: Enabled, Disabled This is a setting related to the CPU's side of the chipset, but it still involves RAM. When the CPU switches from reading to writing, it has to delay. This shortens that delay. Enable it for best performance; disable if it causes problems. You know how it is with these things. Options: Enable, Disable Even if the chipset doesn't detect four banks, this will insist that the chipset uses four-way interleave. If the chipset gets it wrong, the user obviously knows best...right? You should really disable this. The chipset is likely to turn four-way interleave on if you specify it in the Bank Interleave setting. Options: Low, High With a high RAM load (as mentioned in Delay DRAM Read Latch), the signal strength may be insufficient. Change from Low to High to remedy this. Also, this increases stability when using overclocked RAM. This setting only affects stability and not performance. Options: Enable, Disable The chipset can issue read and write commands out of order from the chipset's R-A-W buffer, resulting in higher performance if this feature is enabled. The buffer also has other benefits, so it's a good idea to enable this. Options: 0, 1 When something reads from RAM, the chipset services the request. However, it usually holds the data for one cycle before making it available. Disabling this delay (set to zero cycles) helps performance, but the data could arrive too early and the requesting device could not be ready for it, which would result in instability. Options: 7.8 µsec, 15.6 µsec, 31.2 µsec, 64 µsec, 128 µsec Normally SDRAM and DDR are refreshed every 64ms. However, refreshing every cell simultaneously will result in a power surge. That's not good. So the refreshes can be staggered from one row to the next. 128Mbit and smaller DRAMs like to have this at 15.6 microseconds. 256Mbit DRAMs have twice the rows, so half the interval, 7.8 microseconds, is appropriate. The JEDEC standards do call for 64ms (not µsec!) but today's DRAM can handle more than that between refreshes, so for performance and power (mobile users listen well) reasons, you may want to increase this all the way to 128 microseconds to add a small delay on top of the 64ms already. It helps performance by keeping RAM available for longer. It helps power usage by not refreshing as often. Options: 3, 4 Yet another setting that's faster lower. Set it to 3 if you have badass ninja RAM. Set it to 4 if 3 doesn't work. Options: Disabled, 0 Cycle, 8 Cycles, 12 Cycles, 16 Cycles, 24 Cycles, 32 Cycles, 48 Cycles (Note: cycles may also be called "ticks") This tells the SDRAM how long it should idle before recharging. Increased values allow the SDRAM to postpone a charge and reduce RAM latency. Tune this with the Refresh Interval to get the best settings. I recommend 12 cycles for machines with less than 512MB in total and 32 cycles for all others. Options: Enable, Disable This determines whether the chipset or the RAM controls refreshing. It's better to let the RAM do it itself (enabled), but doing so can cause stability issues with large amounts of RAM installed or with poor-quality RAM. In those cases, you should disable it for a small performance loss but better stability. Options: Enable, Disable Disable this. You don't want to be wasting the L2 cache on fast video RAM when you have slow system RAM to deal with. The bandwidth of your system RAM is unlikely to be over 3GB/sec, yet video RAM can easily top 10GB/sec and a 4x AGP bus is 533MB/sec. The tiny amounts of L2 we get these days to go with our massive system RAM sizes relegate this setting to being Disabled. NOTE: Set to enabled for older systems with older video cards (Mach64, Trio cards, etc.) as they will benefit from the caching ability. If the machine has AGP, disable this. This brings us to the PCI and AGP bus settings, with an ISA chaser. These will mostly be found under "Advanced Chipset Settings," although a few are dotted around elsewhere in your BIOS setup. The PCI specification allows for some variance in it's behavior and certain settings will benefit some systems but not benefit others. Optimizing these settings mostly involve tuning between high bandwidth and low latency; what you give to one, you take from the other. Audiophiles, for instance, are going to disregard bandwidth and have the absolute lowest latency possible. Video freaks will do the opposite, as will hardcore RAID guys. Other settings control how the PCI bus interacts with the ISA bridge and with the CPU. Last and pretty much least, we get to configure the AGP bus. Lots of hot air has been expended debating the importance of these settings, but for the majority of users, they won't have an appreciable effect on your framerate or your benchmark scores. It is recommended that you select the highest performance settings that you can, but if you can't, don't feel too bad about it. Let's go. Options: N/A, 1, 2, 3, 4, 5, 6, 7, 8 This does nothing if you have no ISA cards. The PCI-ISA bridge has a fast bus (PCI) and a slow bus (ISA) to deal with. Transactions with PCI have to be delayed so the ISA side can cope. Usually, the delay is 3.5 cycles. This option allows you to add to that, so selecting 1 will give a 4.5 cycle delay. N/A will leave the setting at default 3.5 cycles. This is an option for overclocked PCI buses as extra delays could be added to keep the ISA bus well synchronized. Options: N/A, 1, 2, 3, 4 See "8-bit I/O Recovery Time." It's the same thing, but for 16-bit transactions. Options: Enable, Disable These enable or disable AGP 2x or 4x, respectively. Enable whichever one your BIOS presents you with. Even if your card can't do AGP 4x, the chipset will fall back to AGP 2x even if this setting is trying to enable 4x. This really is a no-lose situation. Options: Disable, 8MB, 16MB, 32MB, 64MB, 128MB, 256MB AGP allows cards to retrieve textures from system RAM (AGP Texturing). It's very slow and would result in slideshow rendering, but it can be done. Consider AGP Texturing as overflow video RAM; AGP Aperture controls how much system RAM it is allowed to use. The graphics card's faster onboard RAM is normally used for textures, so most users will not need to utilize AGP Texturing. Set the AGP Aperture to 64MB or 128MB and don't expect any performance difference at all from tweaking it, unless you use absolutely massive textures. You'd know if you dealt with such large textures, and games don't. This setting will not alter your FPS or 3DMark score. Options: Various ratios This takes the FSB and chops it up to make the AGP clock. Normally, you would want the AGP clock to be 66MHz, so divide the FSB accordingly. style="mso-spacerun: yes"> On <800MHz Celerons (66MHz FSB), you should have this at 1/1. On 100MHz FSB CPUs like Durons, some Athlons, and some P3s, it needs to be 2/3. Athlon XPs, some Athlons, and other P3s use a 133MHz FSB, so you should set it to 1/2. Newer Pentium 4s use a 133MHz FSB while older ones ran on a quad-pumped 100MHz FSB. You don't want that AGP bus running at something stupid like 100MHz or 133MHz. Some systems have a 1/3 divider. Feel free to crank that FSB up to 166MHz and set the AGP divider to 1/3. (Yikes!) Usually, the PCI clock is taken from the AGP clock so if you're in spec with the AGP then the PCI will also be in spec. PCI runs at exactly half the frequency of AGP. The top AGP clock you should expect to work with all cards is 70MHz. By the time we reach 85MHz, hardly anything is going to work since PCI would be running at 43MHz, 10MHz above the 33MHz spec. Here's some useless historical trivia: Cyrix MII chips sometimes use 83MHz bus speeds. Guess what speed the AGP is? Yep, 83MHz (well, sometimes it's 2/3 and 56MHz) and the PCI clock is out of spec as well. Don't expect it to work. The PCI divider is also like this setting, and the same logic applies, but you're aiming for a 33MHz bus and not a 66MHz one. Options: Auto, 00-FF (hex) There was a problem with KX133, KT133 and AMD751 chipsets with Matrox G400 series and nVidia Geforce and Geforce2 cards. The chipset could not properly work out the AGP Drive Strength (it should have been EA) and so used the standard DA, resulting in crashes or just plain incompatibility. With one of the aforementioned video chipsets (oddly, not the GF2MX), this should be EA. With anything else, it should be Auto or DA, unless you're advised otherwise. Also by raising the drive strength, an overclocked AGP bus can remain more stable, so many use EA or even FA when running the AGP way out of spec. This is *not* a performance-related option and will do nothing for performance. What it can do, however, is damage the AGP card! Be very careful if you choose to increase this above E2. Options: Enable, Disable With these disabled, the AGP waits two cycles before reading or writing. Enable them, and AGP only waits one cycle. When this setting is not supported, there can be display problems with 3D scenes such as incorrect texture mapping, missing polygons, and pixel popping. Options: Enable, Disable This tells the BIOS that your display card needs an IRQ. It's only useful for operating systems that do not properly support ACPI like Windows 95/98/ME and has absolutely no effect on an ACPI (Windows 2000 and XP only, although 98SE and ME have a try) system but it can't hurt to simply set it to "enabled" since all of today's 3D accelerators require IRQs. It also has no effect when PnP OS Installed is set to "disabled," since the BIOS will enumerate the cards and set up the PIC itself. Options: Enabled, Disabled This does to PCI what USWC does for AGP. Since PCI is a bus rather than the port that AGP is, this helps to increase bandwidth. However, it also increases latency across the bus, so there are those rare cards that may crash with this enabled, notably NICs, but most NICs alter the latency for themselves anyway, so they won't be affected. Options: Enable, Disable When the CPU comes to write to the PCI bus, it either does so through this buffer or it doesn't. If it doesn't, it has to contend with anything else that's using the PCI bus because it has to write directly to the bus. If it has this buffer, the CPU can write a quadword then get on with better things. Enable it for better performance, but it may cause instability on systems running a Sound Blaster Live on a VIA chipset, and I can't rightly isolate why. Options: Enable, Disable Allows the AGP device to write to the bus without calling an interrupt. In fact, it's forced behavior, unlike the 2x or 4x modes which are always agreed upon by the chipset and the video card. So if you enable this for the extremely slight performance boost, you'll get instability within any 3D accelerated application when using a video card that doesn't support Fast Writes. Options: Enabled, Disabled This combines the 8-bit and 16-bit settings into one. It seems to add four cycles to both of them. If your PCI is a bit out of spec and you're noticing problems with your legacy ports or any ISA cards you have, then enabling this will help. Normally, it's to be disabled. Options: 1 PCI, 2 PCI, 3 PCI This controls when the CPU accesses the PCI bus. 1 PCI means that the CPU will be granted access as soon as the current transfer is over, regardless of what else is waiting for PCI attention. 2 PCI means the CPU will wait for the current transfer and the one after it. 3 PCI means the CPU will wait for the current transfer and two after it before being allowed on the bus. Delaying CPU access will help RAID and SCSI cards but will hinder TV tuner and NICs. Anything that utilizes DMA will potentially gain from higher settings, but this will slow devices that run through the CPU, since PCI-RAM and PCI-PCI performance is maximized at higher settings and CPU-PCI performance is maximized at lower settings. The CPU will always get the bus within at least three time slices, regardless. Options: Enable, Disable PCI can only be used by one device at a time. If this device is the PCI-ISA Bridge, we pay it special attention (as seen with the Recovery Times), and this is another setting to help. The Passive Release Buffer allows ISA devices to write without needing the PCI bus from start to finish, which would take quite some time.Instead the data is written to a buffer, then transferred up the bus at full PCI speeds. Enable this for best performance, but disable it if your ISA cards, especially ISA SCSI cards or NICs, don't like it. Options: Enable, Disable This only means anything if the CPU to PCI Write Buffer is enabled. When the buffer tries to write to the bus, it may fail. Whether the write is tried again or sent back to arbitration depends on this setting. Enable this, and the buffer will retry. If this option is disabled, the buffer will mark the transaction as failed, and the CPU will send again. Enable this for best performance. However, if your PCI bus is heavily loaded in terms of bandwidth-using devices, you should disable this, as many retries will clog up the bus. Options: Enable, Disable PCI 2.1 specification mandates that this is *always* to be enabled. Strangely enough, it's disabled by default on far too many boards. This causes an ISA device--and you do still have some, even without ISA slots or cards, since the SuperIO chip does a lot of things on the ISA bus--to be told to wait if the PCI bus is in use. It's very much like Passive Release in the regard that it's a buffer between fast PCI and slow ISA. ISA devices now use the PCI bus, with the PCI-ISA bridge translating the data. Since the ISA bus is so slow, telling them to wait is the best idea and it's not going to cause any delay to ISA transactions since PCI is so much faster. HOWEVER, our good friend, the Sound Blaster Live, deserves special attention. The Live appears to not query the arbiter as to the status of "bus parking," and it is hypothesized that it incorrectly assumes a "last device" schema is in use, which is the default in most chipsets. For performance reasons, VIA always parks the bus on the arbiter, which results in faster switching between devices but higher latency for any single device. This option, if enabled, can cause SB Live cards to cause crashes on VIA chipset boards and performance penalties (including high sound latency) on Intel chipset boards. A VIA chipset always parks the bus on the arbiter, as previously mentioned, while an Intel chipset (440BX or better) will park it on the last device to have used it most of the time. This can also affect other cards, such as the Aureal Vortex2, but Aureal patched this up in later driver releases. It's only a real problem in the 32-bit environment of Windows 2000 and Windows XP with ACPI in use. When older methods of assigning IRQs are in use, the cards signal to the arbiter in a different way, bypassing this problem. Options: Enable, Disable If this is enabled, a burst buffer is used when PCI devices want to write to the bus. The data is written when the PCI bus is free or when the buffer is full. PCI 2.1 says this has to be enabled, so listen to the official specs and enable it. It'll make it so that bursting cards don't have to immediately pull an interrupt across the bus. Options: 16 - 255 This controls how long a PCI device is allowed to use the bus. However, it's a minimum. Cards can override this with their own value. NICs often set this to 255 for themselves since a NIC is quite slow and needs a higher latency or it'll be interrupting too much. A higher value means devices have to wait longer to get the PCI bus, but can use it for longer (and are forced to use it longer, even if they don't need to) and so this potentially increases available bandwidth for really heavy devices. The exact setting of this will depend on your particular system, but it's usually recommended to have it set to 32 or 64. Options: Enable, Disable Enable this and PCI busmasters don't have to wait a cycle before hitting the bus. That makes them faster since it removes a 30ns write delay. Disabling it will improve the overclockablility of the PCI bus so if your SCSI or NIC doesn't like to play with your 40MHz PCI bus, try disabling this. Options: Enable, Disable This allows the CPU's L2 cache-to-cache PCI master reads. Good? No. There isn't as much L2 cache these days as there used to be, since we now have much more RAM to contend with. PCI devices may benefit from this, but the CPU probably won't. The jury is still out, but the judge is expecting a "Disable" verdict. Options: Enabled, Disabled Enable this for better performance with a PCI video card or anything else with a framebuffer, such as DVD hardware decoders or TV tuners. Disable it for anything else. This is often a cause of system instability if enabled on an overclocked PCI bus or even just with cards that don't like it. Options: Enabled, Disabled Allows the AGP to send addresses down different lines to the data. Enabled is faster. Options: Enable/0.25% (two options, same), 0.5%, Smart. This compromises system stability by performing frequency modulation on the system clock. Timing critical devices, such as NICs and SCSI may have problems. It's there to reduce EMI problems by making sure the system isn't using any one frequency for too long. Disable it. On a 1GHz system, 0.25% will result in a 25MHz fluctuation in CPU core frequency. Not nice. Always, always, always disable this when overclocking. If you have the Smart setting, the system will shut off the clock to unused slots, so use this if you have it. Options: Enable, Disable This stands for "Uncachable Speculative Write Combining" and has an unpredictable effect on performance. It combines small writes into 64-bit transfers, making more effective use of the bus. However, those smaller writes have to be accumulated first, which delays them. Benchmarking your system is the best way to determine whether you should enable this option. Options: Enable, Disable This is for use with ISA VGA cards with VESA feature connectors that are in use. It prevents the overlay from being corrupted with an incorrectly mapped palette. TV-tuners and MPEG video cards are most commonly affected but here's the big catch: It reduces video performance and is also useless for PCI or AGP displays. Note that some older PCI cards may still use the ISA address space and so would have their overlay problems fixed by this. Much the same way, some early AGP cards (SiS 6235 and ATI RageIIc for example) weren't AGP cards at all, just PCI cards that fit into an AGP slot. Options: 0, 1, 2, 3 This determines the signal strength on the FSB--higher value, stronger signal. This can reduce chipset stability as it will increase chipset heat production, but it will also increase overclocked CPU stability. Set it to 2 or 3 when overclocking, and set it to 1 when you're leaving the CPU at its marked speed. Like other drive strengths, this is a stability option and will do nothing for performance. Options: Enable, Disable This enables the control "S2K Strength Value." Disable it. You don't need to change the S2K Strength unless, perhaps, you like the look of white text on a blue background, the first word being "STOP." Options: auto, 1, 2, 3, 4, 5, 6, 7 S2K is AMD's name for the Athlon's bus. On Athlon Classic, this also covered the Back Side Bus (L2), so it was separate from "EV6." This setting fine tunes the signal strength of the bus itself and has only been seen on Asus A7V series motherboards so far. This is an absolute black art. Alter it if you know precisely what you're doing and have an in-depth knowledge of both the chipset and EV6 buses.Alter it at your peril. Most people should leave it either disabled or on auto. Options: Various hex values This is similar to caching, but instead of polluting the L2 cache, it uses a bit of the UMB of system memory where nothing else can really go. System RAM is faster than the BIOS EEPROM so BIOS calls will be faster. The memory ranges are for shadowing PCI address space. If your PCI card maps its BIOS into one of these ranges then you can shadow that too. This is useful for SCSI cards running under Windows 3.x or MS-DOS. Shadowing a BIOS has no effect after the HAL (NT) or drivers (9x) are loaded, since these do not call the BIOS at all and instead directly interface with the hardware. WARNING: DO NOT SHADOW THE BIOS WHILE TRYING TO FLASH IT. THE BIOS WILL NOT BE COMPLETELY WRITTEN. THIS APPLIES TO VIDEO BIOS OR ANY OTHER SHADOWED BIOS. What happens is that the flash program writes to the shadow RAM and then when this area is fully written to it will write the rest of the data to BIOS ROM. So if we have a 128kB BIOS in a 96kB shadow, then the first 96kB of the BIOS will not be altered at all, but the last 32kB will. You don't need me to tell you that the BIOS will never work again. More recent flash tools from Award and Ami will attempt to detect a shadowed BIOS and either abort or attempt to override it. Options: Enable, Disable This refreshes the ESCD data on every boot, regardless of whether it's changed or not. I can't think of when this would be actually useful, but sometimes the ESCD can become a little confused if you've moved or swapped cards around. Options: CHS, LBA, Large/Other CHS is for drives smaller than 512MB. LBA is for drives that are larger. "Other" or "Large" are for drives that do not support LBA after 8GB, usually very old drives that are larger than 8GB. Options: Enable, Disable Some tell us to disable this under NT4. Some tell us that NT4 is fine with it. It appears to be service pack and chipset dependent. So, if you have a 440BX or lower, or a Socket7 chipset, then disable this. Also disable it if you are running NT4 with less than SP5. Options: Enable, Disable Enable this for Windows 95, Windows 98, Windows NT4, Windows ME, as well as Linux kernels that do not support ACPI.Disable this for Windows 2000 and Windows XP as they use ACPI and do not require the BIOS to interfere with the PIC in any way and will reset any configuration data in there anyway. Microsoft recommends this option to be disabled when ACPI is in use, since the PIC/APIC is used in a different way under ACPI. I've tried both settings with an ACPI Windows 2000 system, and I couldn't tell any difference. Options: 1, 3 Legacy ports use the ISA bus; this determines which legacy "DMA" channel the ECP parallel port will use. It's usually set to "3" and the setting is only a real concern if you're using ISA sound or network cards that may conflict with it. Ancient ISA cards set their DMA with jumpers. When selecting the DMA channel for the parallel port, we need to make sure it's not using the same DMA as an ancient ISA card. By ancient, I really mean it--we're talking before 486 here. Many people do not use their parallel port and can disable it, but some do, mostly for older printers or zip drives. style="mso-spacerun: yes"> Most new printers use USB connections, but some still use parallel ports. Anybody with a parallel port printer should have everything on Auto. These options should remain at default unless you know that you have to change them, or you wish to disable the ports. Options: Enabled, Disabled This option has so many names it's not funny. All it does it allow a USB keyboard to work in Win 3.x, MS-DOS, and other environments that are not USB aware. CAUTION: On my KT7, this option kills a USB mouse after resuming from Hibernate in Windows 2000. I recommend you disable this unless you need it to use a USB keyboard for, for example, flashing your BIOS from a DOS boot disk. That's it. While your own motherboard may feature a few specific options not covered here, hopefully you've learned a bit about the thought process behind configuring a motherboard's BIOS options and the trade-offs involved. Also, if you've done your research and you still can't find what you need, the Ars CPU & Motherboard Technologia forum is a great resource (but I mean what we say at the beginning of this sentence). When you are faced with an unstable system, remember to first check your connections, card placement, and BIOS settings before assuming you have faulty hardware. Poor BIOS settings can often be the cause of a slow or unstable system. By taking the time to carefully configure your system, you'll have a system that both performs better and is more stable. |
||||
