Intel LGA (land grid array) used to connect processors to motherboards is being replaced by BGA (ball grid array) which welds the processors to the motherboards, with no ways back to future upgrades of processors, and the future buyers will have no options for processors upgrades unless change the whole kit (mobo and processors). I had a problem whith LGA motherboard connectors, I bent one or more mobo contacs while fixing the processor to the motherboard and had to buy a new one, and new glasses as well, as I blamed it to my lack of vision accuracy for small things at short distances, which I wasn't aware of.

I understand. I also keep an eye on legacy PCI slots. My searching is more restrictive: DVI, PS2 (I refuse to stop using my Northgate keyboard), 4th generation Intel, more than 4 SATA ports.

Interestingly, the boards designed by Intel do not have PS2 ports. Many bottom-of-the-barrel boards too.

Looks like a candidate for my list.

Yes and no. I've been hitting the web hard browsing the specs of motherboards. Many boards have PCIe but not always have legacy PCI slots.

All benchmarks have their limitations and I didn't mean to promote Passmark as anything special. Since someone was making a hard general claim that "Intel CPUs are terribly overpriced" "Intel chipsets are unreliable", and that the "AMD 9590 is a bargain compared to the Intel i7-4960", as if this was just accepted fact, all I had to do was to find ONE benchmark to show otherwise to discredit that approach. If Reaper is actually interested in finding out how deeply biased he is, he can look up several. I'm betting not one single benchmark ANYWHERE, by ANYBODY, shows the AMD-9590 keeping up, let alone trouncing the Intel i7-4960.

Exactly my point. It stood head and shoulders above all competition, including it's own, for many years.. not exactly an example of a company incapable of making a reliable chipset. The bottom line is that no matter how big a company is, a smart buyer researches the exact product and leans away from either endorsing or damning all products by said company. Influenced by track record? Sure! Convinced? Never!

It never ceases to amaze me how many posts are up on the net that go something like "I bought a Western Digital (or, insert manufacturers name here) hard drive and only had it 6 months and it broke! Never buy Western Digital (or /insert) ever again. They suck!" These people have no idea how amazing it is that hard drives work at all and more to the point the absolutely incredible low return rate... less than 1/10th of 1 Percent iirc. Sure if it happens to you it feels like 100% but it isn't justification for writing off the other 99.999%.

to be clear, I don't hate AMD, quite the opposite. When I discovered they had headhunted a lot of the boys over at the DEC Alpha project and read about the Athlon, I bought stock as well as several CPUs. The Athlons are paperweights now, but I'm still using the FX-57 and when AMD hits another breakthrough, I will buy again. Right now, it seems Intel owns the ballgame, although they need to go to school on Graphics. There, AMD possibly has a leg up.

This is only true for their mobile department, the desktop and server chips will keep the LGA interface.

PS/2 to USB adapters are quite common and cheap (around 5$ as it seems), so you shouldn't limit your choice by that.

Perhaps, yet a perusal of the reviews of such adapters indicate most don't work because they are cheap pass-throughs. Even if I found an adapter rated five stars, the Northgate keyboard is actually a 5-pin DIN. I use a DIN-to-PS2 pigtail to convert to PS2. Whereas USB is designed to be hot pluggable, not so much for 5-pin DIN and even if true, I'll never take the chance.

Many modern motherboards still provide PS2 connectors. Apparently I'm not the only person on the planet refusing to go completely USB.

Usually gamers are those that want to have PS/2, since USB keyboards do not support N-Key-Rollover, but PS/2 does. So maybe you should look for boards aimed at gamers.

EDIT: For example, the ASUS Maximus IV series and the Gigabyte G1.Sniper series.

The NForce 980SLI wasn't too bad, but yes, it has lagged behind the times as of late. The NForce 4/4x0 based chipsets were the worst, as I know first hand with my NForce 430 motherboard in my home PC. It's not bad, but it's not great.

SiS... yes, they haven't produced a chipset in ages, but damn they had some serious stability and reliability. I actually wish they were making chipsets again. They were great.

For VIA actually I could take you back to the KT133 series in terms of terrible. They had one of the worst PCI and AGP implementations available that drivers required numerous patches to get stuff stable enough to work with their protocols.

For Intel, just about each chipset I've had burnt out within a year. The only one that lasted longer was an old 82810 chipset that was next to useless due to the lack of a AGP slot. The Intel 900 IGP based board I got later on burned out within a year, an 82830 I had before that only lasted 6 months before it tanked, and five clients I service PCs for all have had motherboards replaced within the last year and most were newer chipsets also.

Intel chipsets I used that are still alive:
- i915, ASUS EeePC 701, I use it as ebook-reader
- i945, Intel mainboard with an Atom CPU, running 24/7 for about 4 years now, no problems at all
- P45 on a Gigabyte GA-EP45-UD3, managed to overclock my C2Q 9550 to stable 4GHz, this was a really good one

None of them has any signs of "burnout". From my experience, it is very rare that it is actually the chipset that goes down the drain when a motherboard dies, usually it are the voltage regulators and capacitors.

ReaperX7, which brands where those mainboards and are you sure that actually the chipset and not some other component failed?

The DIN/Mini-DIN keyboards use the same simple TTL protocol first used by the IBM AT. It's not designed to be hotplugged, and doing it can damage your mainboard irreversibly.

For almost a decade I used a IBM Model M keyboard together with a DIN/Mini-DIN adapter and a PS/2-USB converter. I didn't have any issues.

Originally I did go USB early, because I didn't want to risk damaging/wearing out the MiniDIN(PS/2) port by the weight of the heavy DIN adapter, so I connected the construction to a cheap USB converter and the result to the more rigid (and less important/replaceable) USB 1.1 port instead. Later I used the USB 1.1 hub integrated into the base of my CRT monitor.

Today I have a USB keyboard, which is basically an AT MF2 keyboard with an internal converter. I also directly switched from a serial (RS232) mouse to USB and never used PS/2 mice.

There are two common kinds of failure being attributed to chipset.

1) Longevity vs/ Early Burnout - and of course this is generally a power problem - bad regulation, poor quality capacitors, etc. not really a chipset issue. (Do some cheaper power supplies still bypass Power Good?)

2) Stability & Function - Chipset is definitely a factor here but so is the support system

Regarding VIA - While I do agree they have a spotty track record, I did buy 7 Super Micro Super 370SDA boards at $20 US each, that utilized the VIA Apllo Pro 266 Chipset. They came out before Tualatin and required modification to run those but still supported up to 3 GB DDR Ram at a time when all were still using PC133 and limited to 512MB. All 7 still run and I used 2 of them daily until 3 months ago when I moved residence. Those 2 operated a Minecraft server up 24/7 for over 3 years.

Regarding Intel - I have an Asus P3BF w/ 440 BX chipset that I dragged into the future with a PCI IDE accelerator card and a Slotkit. It was my main machine up 24/7 for many years. It still runs but what's the point anymore? I just haven't the heart yet to throw it out.

I have an Asus TUSL2 board w/ i815 Chipset that was superbly stable and also still runs as of 3 months ago but again, stuff that old is severely ram limited.

Maybe I've just been lucky or maybe it's because I build a clean system with top-notch power supplies, but I have never had a burnout, ever... even though I tend to overclock. The closest I came was an nForce 3 MSI board upon which USB ceased to function but apparently that was somewhat common. However it is also possible that I've been lucky because I'm not an early adopter (usually) and wait till I see some feedback.

It seems Woodsman knows exactly what he wants and is willing to research to get it. I'm betting he will do just fine

PS Although this lacks some features Woodsman wants it is interesting that Super Micro has made a foray into the DIY user. Looks like real capacitors http://www.supermicro.com/products/m.../c7z87-oce.cfm

The i915, which I'm pretty sure is what it was as it was paired with a Pentium D, was a Dell re-branded Intel motherboard I bought with an XPS system years ago. I no longer have it so any other details would be spotty, but the Intel southbridge chipset completely died. No capacitors or anything popped or were damaged. The chip was bad after a RMA to Dell confirmed it.

I forget the other one with the 82830 completely, but remember it actually had a serious hardware fault which fried the Northbridge, Southbridge, my RAM modules, and my hard drive. The board was recalled by the OEM eventually, but I can't remember the brand of the top of my head.

The old Toshiba rebranded 82810 board is still working. It's slow as a slug in winter with a paltry Celeron 533MHz CPU and 256MB RAM, but it still works as a file backup server. I even still have an old Voodoo3 3000 PCI, an Intel PRO 100 PCI Ethernet card, and even the onboard Aureal 8810 Vortex Advantage all still work. Even still has a 60GB HDD IDE drive with Slackware-Current 32-bit installed.

My SiS 730S board died only due to a power-supply failing and damaging the capacitors and CPU. That little sucker ran for 7 and a half years with Windows 2000 and XP in combination with Slackware Linux and it worked for a long time beautifully.

Peace on earth and good will to all. Once upon a time I did a lot of research how to achieve that, but I'm not an ivory towered pinhead, movie celebrity, or big star athlete, thus nobody listened to what I discovered. Nowadays I just tinker with computers and keep wood in the fire to stay warm.

The thread has gone astray but my original two questions were answered:

* Don't worry about UEFI and enjoy all the new features.

* Integrated Intel graphics will do fine for my needs. (Not to mention that 4th generation Intel APUs/CPUs are fast and energy efficient.)

Woodsman- yeah, intel video should work. But if you are buying a i7 I'd get a $40-60 video card, if only to save on CPU and RAM I/Os. Also, when you use the intel video, it uses the CPU L3 cache.

I hope you dont mind the following, just I really like this sort of discussion.

I'm yet to see a reliability benchmark, and considering that a i7-4960X is $1K+ US, where a FX-9590 is about $400, I'd say that intel CPUs can be (not always) overpriced, and the AMD is a bargin...in some ways.

Challenge accepted!

http://vr-zone.com/articles/ivy-brid...s/54295.html/2

Bottom table, maya-03.

Yes, the i7 4960 is faster almost everywhere....at well over double the price, you'd expect that. Its not like the old days when a $1K+ P4EE would go neck a neck with a top of the line AMD for $350-400....

A fairer comparison is FX-9590 vs i7 3770/4770, or FX-8350 vs i5 3XXX/4XXX.

The 1GB+ RAM limit of the VIA chipsets is what made the step back from 1GB max RAM with the BX to 512MB max RAM with the i810/i815 look very cynical. Which is was. Remember that the BX chipset was used with boards like the Abit BP6 and dual celerons. Intel disabled the SMP on celerons post 533MHz to stop people from running 'budget' CPUs in 'server' sertups.

Also, the VIA athlon chipsets didnt have intergrated video (apart from the KLE133), which was always looked down on by many poeple. Intergrating the failure i740 video card into the chipset was also pretty dodgy.

If AMD hadnt came along with the athlon, intel would have closed things up much sooner. IMO the was the end goal of itanium was all 'top end' systems, not just servers. AMD really scewed that one up.

But IMO you'll never see AMD ctach intel again in the top performance crown. Intel made a huge list of mistakes, misteps, and just underestimation of the competition that let AMD not just catch intel, but pass them for a few years in the athlon/early athlon 64 years.

Lucky that your old VIA boards ran that long, most of the VIA DDR P3 boards I've heard of are dead. Or maybe its not totally luck.

Wow. Just wow. I've got a dell i810 system I found by the side of the road that still runs. Theres an asus i815 board somewhere round here that was running 6 months ago, and I'ev got quite a few old P4s I need to get rid of with i845 chipsets that still run, my housemate is now using a i915 chipset P4D that I was given, and so on.....

Look, this is totally unscientific opinion, but in my experience some people just have issues with hardware from some companies.

My brother in law went though 4 (yes four, f-o-u-r, 4) A7V 333s in 3 weeks, then some A7V400 variant. Everything died. Even the CPU was eventually found to have some ultra rare manufacturing flaw. He had no issues with intel before or after that.

I have no idea why that should happen, but it does, and its not the only example of this sort of issue I've run across. I'd avoid intel like the plauge if I were you. Not that intel makes that much really bad stuff. Even though you obviously have issues with the hardware.

KT133 was pretty nasty a 'tweaky/unstable' chipset in many ways. I've always suggested avoiding them, but I didnt lay as much of the blame on the KT133 on VIA as I do with later chipsets.

Looking back at the asus boards of the start of the DDR era (using asus because they are the ones I had the most experience with, and I know many people who had the boards as well). We had-

A7V 266/333. VIA. Not bad when they ran, but awful relability IMO.
A7A 266. Ali. Only ever saw a couple, they semed to last almost as well as the SiS/nVidia. Odd, because the SD RAM/DDR setup would have made upgrading easier.
A7N 266. nVidia. Tough. I have one that still runs that I pulled out of a plactic bag on the floor of a music shop where I used to work.
A7S 333. SiS. Amazingly tough for what they were. Not that fast, but better running than dead.
A7M 266. AMD. Toughest DDR athlon chipset ever.

OH SURE! Nice job phonying up that website and making those "graphs".......Prolly did it on an Intel machine too just for irony.......

Oh wait, heheh, just kidding. That was actually an interesting comparison. While a pretty wild anomaly at first glance, later comparisons show that as graphics is notched up they seem to nearly merge. It appears that if you can live with the power consumption and heat (and a fella can buy a lotta fans and a fair bit of current for ~$500 :P ), the FX-8350 is indeed somewhat of a bargain.

Somewhere in the midst of us there is the Truth that brands and models exist for a reason and that reason can be different for different people and jobs. While I don't think I could ever stop researching products before I buy (wouldn't something, I dunno... just snap!?... and.. and...let the smoke out? heheh) The Door Prize often just goes to the luckiest.

In case of the FX9xxx series this could indeed happen. Most common boards can't handle a 200W CPU and will indeed let the smoke out.