AMD or intel - reliability?

locutus12 · 18 Jul 2006 at 20:54

NathanE said:
It is the chipsets that often let down AMD CPUs. Not AMD's fault though.

VIA

Corasik · 18 Jul 2006 at 21:03

Tom's have certainly been AMD fanboys in the past, I remember just before the 800Mhz FSB P4 northwoods came out, the AMD XP's were slightly faster than P4's in most tests.

Every time the AMDs were 1-5% faster toms would say things like AMD dominates the P4 in this test, or AMD's superb performance show here. Seriously exaggerating such a tiny difference that even results which were subject to 'random deviation' would be claimed as stunning AMD wins.

Yet in the encoding tests where P4 did really well, even if the P4 was 10-15% faster than the amd, the comments were more like 'P4 has a slight lead here', or a 'small win for intel on this test'.

AMD64's was such a good chip in its day that in a lot of tests is really did perform a lot better than P4/Netburst, so 'fanboyism' didnt count for much. Simply reporting the truth was good enough.

As for the topic... Reliability. Well Ive used both systems, and once up and running both are pretty good, but I've found that AMD systems seem to be a lot 'fussier' about ram than P4's. I guess thats because the IMC that AMD use can push the ram with tighter timings, so if the ram isnt 100% the AMD is likely to suffer more than the 'slower' intel/northbridge memory interface.

Now with Core2 even if they are fanboys the chips are reviewing so well, they even fanboys like Mad Mod Mike have had to admit they are decent.

FreeStream · 18 Jul 2006 at 21:32

Holy thread revival! :eek:

Gashman · 18 Jul 2006 at 23:01

intel in sheers indestructability anyday but definately AMD for stability and performance, i mean matt has smithfield, running at about a billion *C once, i mean absolutely roasting, boil potatoes or fry egg sort of temps. quite frankly you could drop a nuclear bomb on an intel and it would keep going, there so resistant to overclockers abuse its unbelievable

CurlyWhirly · 19 Jul 2006 at 14:05

locutus12 said:
VIA

I still have an AMD Athlon XP 3000+ and a VIA chipset on a Gigabyte mobo and I have been running them for over 4 years with no problems apart from a 'crackling' sound caused by PCI latency (which has nothing to do with processors I know) and it was fixed with a patch.

Perhaps I am just lucky

Concorde Rules said:
Holy thread revival!

Yes

NathanE · 19 Jul 2006 at 14:59

CurlyWhirly said:
I still have an AMD Athlon XP 3000+ and a VIA chipset on a Gigabyte mobo and I have been running them for over 4 years with no problems apart from a 'crackling' sound caused by PCI latency (which has nothing to do with processors I know) and it was fixed with a patch.

Perhaps I am just lucky

I think you must be

Can I introduce you to a little friend sitting in my cupboard? His name is "Epox K7XA"!

NathanE · 19 Jul 2006 at 15:00

Gashman said:
intel in sheers indestructability anyday but definately AMD for stability and performance, i mean matt has smithfield, running at about a billion *C once, i mean absolutely roasting, boil potatoes or fry egg sort of temps. quite frankly you could drop a nuclear bomb on an intel and it would keep going, there so resistant to overclockers abuse its unbelievable

Surely indestructability and reliability go hand-in-hand, hence a slightly contradictory post?

drunkenmaster · 19 Jul 2006 at 15:37

people constantly have a go at via, to be perfectly honest, with many ath xp, and couple 754 boards, one 939 i've never had an issue with via. they had lots of issues sure, but 99.9% of those, and nvidia's, and sis's, and amd's AND intel's issues were quirk's/bugs when used not at stock, the question is, so the hell what. no one has to make a board that works anything other than stock, and so far, even the k7s5a(i think that was it) ECS board that had a billion little quirks everyone knew someone who had one run for 5 years without a problem.

reliability when stock is a thing that is virtually assured at 100%, reliability when overclocked is completely different as no one can say for certain what the 100% precise speed their cpu can do overclocked, or exactly what fsb, timings their cpu/mem/mobo can do so often you'll think you're stable only to have a crash 2 months later in some new game that pushes your rig harder than ever before. this doesn't make any part of the system unreliable it just means you didn't know you'd overclocked to far before.

quite frankly, when it comes to sis, they had the first proper desktop aimed dual channel p4 boards out, and it was a freaking fantastic board, solid, clocked to high hell and easily spanked any other p4 board as it was dual channel.

matt100 · 19 Jul 2006 at 17:38

if gfx cards and memory were as reliable as ANY of the cpus I've had I would be a very happy camper..

Gashman · 19 Jul 2006 at 19:09

NathanE said:
Surely indestructability and reliability go hand-in-hand, hence a slightly contradictory post?

nope not at all, i deem reliability to be stability as well as reliable life-expectancy wise, thing is, intels stability is proberbly directly related to heat, so get a stonking cooler and it'll be stable as a continent and nigh indestructable as well

Cyber-Mav · 19 Jul 2006 at 19:12

cpus are just as reliable as each other, its the chipsets that distinguish the brands from one another. intel seem to have more reliable chipsets in the past hence why intel had the feel of being more stable than amd.

NathanE · 19 Jul 2006 at 19:21

Gashman said:
nope not at all, i deem reliability to be stability as well as reliable life-expectancy wise, thing is, intels stability is proberbly directly related to heat, so get a stonking cooler and it'll be stable as a continent and nigh indestructable as well

I'm not following. Are you saying Intel Smithfield & Prescott's are less reliable because they run hot?

Gashman · 19 Jul 2006 at 19:32

no but my friend uses stock cooler with smithfield (two prescotts bolted together) running a small amount over stock and he gets a lot of random exiting of games and slowdown which really has to be attributed to his processor, since chipset and GPU temperatures are fine, his CPU runs at like 75+*C when under load and more than mine (at load) when idle, there reliable to the most part but suffer with thermal issues, which is one of the reasons why i chose this venice over a pentium IV, that and performance and potential overclockability

Gashman · 19 Jul 2006 at 19:34

Gashman said:
no but my friend uses stock cooler with smithfield (two prescotts bolted together) running a small amount over stock and he gets a lot of random exiting of games and slowdown which really has to be attributed to his processor, since chipset and GPU temperatures are fine, his CPU runs at like 75+*C when under load and more than mine (at load) when idle, there reliable to the most part but suffer with thermal issues, which is one of the reasons why i chose this venice over a pentium IV, that and performance and potential overclockability

ordered an akasa evo 98 heatpipe cooler yesturday on his behalf from competitor, should arrive here on friday, and we'll see where that takes him

Digital Punk · 19 Jul 2006 at 19:43

Cyber-Mav said:
intel seem to have more reliable chipsets in the past hence why intel had the feel of being more stable than amd.

As someone who's putting together and configuring servers on a daily basis, I can tell you that Intel have their fair share of chipset problems too.

Up until a couple of years ago they relied on Serverworks chipsets for all their server class motherboards.

Intel's chipsets are good, but they have issues just like any other manufacturer.

NathanE · 19 Jul 2006 at 19:56

Gashman said:
no but my friend uses stock cooler with smithfield (two prescotts bolted together) running a small amount over stock and he gets a lot of random exiting of games and slowdown which really has to be attributed to his processor, since chipset and GPU temperatures are fine, his CPU runs at like 75+*C when under load and more than mine (at load) when idle, there reliable to the most part but suffer with thermal issues, which is one of the reasons why i chose this venice over a pentium IV, that and performance and potential overclockability

Tell him to install some better cooling if it's running that hot.

It's more likely to be driver related.

Also you can't base your entire opinion of Intel and reliability on just your friends case.

To be fair though "Smithfield" isn't exactly Intel's most reknowned piece of engineering

Gashman · 19 Jul 2006 at 20:35

lol thats true, but all 'prescott' based things run well warm, celeron Ds, pentium IVs and smithfields, all rather toasty, pentiums aren't reliable with adaquete cooling, AMDs are happy with any sort of cooling

weescott · 19 Jul 2006 at 20:42

It's nothing to do with AMD or Intel being more reliable, it's down to the build and more importantly the end-user!

NathanE · 19 Jul 2006 at 20:50

Gashman said:
lol thats true, but all 'prescott' based things run well warm, celeron Ds, pentium IVs and smithfields, all rather toasty, pentiums aren't reliable with adaquete cooling, AMDs are happy with any sort of cooling

Not true at all. Prescott 478 ran hot, yes, but the 775 Prescott runs just like any P4. P4's in general don't run any hotter than their equivilent AMD's. There is a Prescott 775 sitting upstairs. All it has is the stock Intel cooler and one exhaust fan on the case. Zero effort went into its cooling stategy and yet it runs at idle 32c and load 42c. These temperatures are in no way going to affect the performance or reliability of the system.

Also the statement "AMDs are happy with any sort of cooling"... you and I know that simply isn't true. Look at the FX60 pumping out >130W. It's no better than a Smithfield... Don't forget about the Thunderbird's either.

Also the topic is about reliability not warmth. Warmth and reliability do not correlate if you plot them on a graph. That's not to say that chips don't have a threshold - after which they'll start crashing etc, but up to that point (and a properly built system won't ever reach that limit) warmth is a non-issue.

CurlyWhirly · 20 Jul 2006 at 14:19

NathanE said:
Can I introduce you to a little friend sitting in my cupboard? His name is "Epox K7XA"!

I don't get it.

Are you saying that the K7XA gave you trouble back then :confused: