For any of your benchmarks on CPUs, is there any QUALITATIVE difference between an OEM and Retail chip? In other words, could I expect better performence from a retail than I could from an OEM? I know all about the physical differences, and I plan on setting up my own cooling system (Alpha PAL6035MFC), but I want to know if I will lose any real processor performence if I use OEM instead of the commonly accepted Retail
Thanks So Much,
Jason Brown
my experience: I have not seen any evidence that a retail processor clocks better of faster than OEM. My current cpu's are OEM (Intel and AMD) and they are at the limits of reported speeds v.s. config comparison (voltage and cooling methods).
This isn't to say there isn't - but no-one I have seen has produced a chart of retail v.s. oem overclocks.
-* This Space For Rent *-
email for application details
Me either. There was a rumor going around for a long time that retail was special, but recently I think that's been dismissed. The only reason that retail costs more appears to be for the 3 year warranty and the stock HSF. It's the same thing with other components. You won't suffer one bit buying an OEM cpu, a hard drive, or a sound card etc... You might miss out on some of the bundled software, but that doesn't apply to a CPU so who cares....
<font color=red>Amd or Intel? Who cares?? Not me...</font color=red>
I think it's not so much of a quantitative difference as more an added assurance of quality. From what I've heard (I have never woked at a FAB so I have no evidence to back this up.) the retail chips have passed higher quality assurance testing.
When overclocking there is always the chance that it won't overclock well. Such are the joys of overclocking. It's always a possability that you'll get no performance boost from a chip, however unlikely it actually turns out to be.
OEM chips, theoretically, have a higher chance of not being very overclockable or being slightly unstable when overclocked even at reasonable settings.
Retail chips, theoretically, have a better chance of overclocking without problems.
Again, I have no proof of this, but it sounds reasonable enough. After all, OEMs don't go around overclocking. They just need a chip that runs stable at that speed only. And OEMs almost always cut corners with cheaper parts. Plus their waranty on the chip is much shorter. So it all stands to reason that the OEM chips are simply cheaper, in every sense of the word.
As the old addage says, you get what you pay for.
If the opposite of pro is con, what is the opposite of productivity? Ground first.
Not sure I agree with your logic, but I won't say you are wrong.
Let's say you are Fred the foreman in charge of binning cpus off the production line for cpu company X. Your product line is (simplified) 1.4G, 1.2G, 1G and .8G.
Now, do you test chips that all roll off the same plant (no special manufacturing process for retail v.s. OEM) to see if they'll do 1.6G? Probably not since you are not selling 1.6G cpus yet. Marketing may do dsamples to see if yields are high enough to start the range though.
If I were binning I'd start at one end and work down or up, depending on what your yield tended to produce. You test at 1.4 - stable, bin it. Not stable, test at 1.2, not stable test at 1.0.....
What would this mean? Well if all chips are the same (I don't think that manufactuters deliberately now try to make 800Mhz cpus, just some of their yeild will statistically be flawed to only give this even though they are aiming for 1G.
Assuming they can fill needs for higher clocked cpus, the excess starts rolling down the bins to the lower clocks to fill demand, even though they are capable of more. The chips at the top have all passed their 1.4 test, or whatever and get split oem and retail.
Now - if I was picky, id say the rolled down higher capable cpus only went into the retail bin, but who knows if this is true? I guess not since now we can buy OEM 1G AXIA that does 1.4G or 1.5G.
It does not seem, in the proof, that AMD have any really bad chips rolling out lately. That of course is not possible, but maybe their 1.4 yield is so high that everything for the lower speeds is actually overflow from the 1.4 bin and therefore capable of much.
Now this is all pure speculation, and it's late at night so my thoughts may have wandered somewhat, but I don't think there are many bad new chips out there right now since current architecture has done so well the yieds are much higher....
-* This Space For Rent *-
email for application details
I'm not saying that there are any bad chips out there, just less overclockable chips.
Say I were a chip manufacturer and I was making a product line of 1.4, 1.2, and .8 GHz chips. My company's retail chips have a 3 year factory waranty and my comany's OEM chips have only a 6 month waranty.
So I knock out a batch of 1.4GHz chips. I run them through my retail testing standards. A quarter of them pass with flying colors. The other 3/4s don't. So I label that 1/4 as 1.4GHz retail. Then I run that 3/4s through less-intensive OEM testing. Of that another 1/4 pass, so I label that 1/4 as 1.4GHz OEM.
Now I still have half of a batch left unlabeled. Even underclocked, I can't entirely trust them to last 3 years if they failed the retail test the first time. So I test them underclocked as 1.2GHz OEM. 90% of those pass that test, so they get the 1.2GHz OEM label stuck on them. The remaining chips get tested at OEM .8GHz, which all pass, so those get labeled as .8GHz OEMs.
This way I ensure that the retail chips with the longest waranty are of the best quality. The OEM chips are likely to last their 6 months without a problem, but possibly only just that. After all, if it dies after that, it's the OEM's problem, not my company's.
The retail and OEM chips at 1.4GHz all operate at 1.4Ghz just fine... for 6 months anyway. However, the retail 1.4GHz is more likely to overclock without problems than the OEM 1.4GHz because it has passed stronger testing. And the retail 1.4GHz is likely to last longer than the OEM 1.4Ghz when overclocked.
Of course, this is all a very general description, but it is how I'd do it if I were a chip manufacturer. Again, I have no proof that it is the way things are done. It just makes sense logically to me.
If the opposite of pro is con, what is the opposite of productivity? Ground first.
Well I can give at least some input on this subject, as I have had both a retail and an OEM athlon. They were both 1.2 ghz 266 mhz FSB chips, with the same steppings no less, AXIA on both. With the retail chip, I was able to run stable at 1.33 GHz ( 10 x 133 ) with no problems, but was not able to up the FSB at that point. I could run a higher FSB if I lowered the multiplier, but it seemed to top out at 1.33 no matter what I did. The OEM chip was a little better. It would run at 140 x 10 for 1.4 ghz. If I lowered the FSB to 133, it would run at 1.466 ( 133 x 11 ). So from what I have seen, retail CPUs are not always better than OEM ones. I just think the retail ones have a warrenty, so they cost a little more. It's like insurance, you have to pay for it.
P.S.
The retail chip, which was the first I bought, is now in the hands of some cold hearted rat bastard theif, who stole my computer but left behind thousands of dollars in home electronics. Go figure.
That oughta void your warrenty!
hey - really sorry to hear about your PC man - all we can say is that it could have been worse I guess. You could hope the heatsink came loose and he fried the cpu, and got no value himself from the system....
-* This Space For Rent *-
email for application details
I don't think that the QA for OEM chips would be lower. The bulk of a good companies sales would aim to be oem. The likes of Dell and Compaq would very rapidly drop a manufacturer that consistently delivered goods that were not reliable. Just because components are OEM, I don't buy that they are lower quality. This makes less sense that saying they are less overclockable. AMD, or whoever, would get reamed by major OEM vendors for supplying crap goods. If anything, the retail sector would be easier to ream as the little guy has less comeback - warranty or not.
Personally I'd also rate any cpu warranty as almost valueless. If it doesn't fail in the first few months vendors know you are not going to return the thing. Some people will, power to them, but if that cpu fails in 2 1/2 years time, the value will be $40 tops and the vendors had your money in the bank. They wouldn't care. I'm not saying it right or wrong, but economics and cosumer habits would just say they'll win out.
-* This Space For Rent *-
email for application details
I'm not saying anything like ALL retail chips are better than OEM chips. I mean most sales are to OEMs, so I'm sure that a lot of retail quality chips get sold to OEMs. And as such, an OEM chip could easily end up out performing a retail chip because they could both technically have been originally retail chips.
(Honest, I'm not trying to make this confusing.)
But my point mainly is that retail chips are meant to last 3 years. OEM chips aren't, and OEMs won't complain so long as the chips last a year or so. So I'd simply put more faith into the quality of a retail chip over an OEM chip.
I mean most OEMs don't give a flying fig about quality; they only care about marketting.
Retail on the other hand is generally dealing with skilled PC technicians who are going to make sure that everything works perfectly and are willing to fight for their PC. As such, quality is more important to the retail sector than the OEM sector.
I could very well be wrong. Maybe the only difference is the packaging. I'd be kind of surprised if that were the case though.
I wish we had someone who worked at a FAB and knew the actual answer to this. You would think at least one member of THG would know for certain.
If the opposite of pro is con, what is the opposite of productivity? Ground first.
| Quote : But my point mainly is that retail chips are meant to last 3 years. OEM chips aren't, and OEMs won't complain so long as the chips last a year or so. So I'd simply put more faith into the quality of a retail chip over an OEM chip.
|
I think this is about 100% incorrect. OEM will certainly complain if all their PCs die within 12 months, their marketing people would go ape because they'd have a PR nightmare and all the companies that bought $1,000,000s of equipment would boycott their products. I think if anything OEMs want, and get, stability and longevity.
-* This Space For Rent *-
email for application details
Actually, I think that you're dead wrong on this one.
OEMs don't care about quality. If something goes wrong, they've got so many replacements in stock that they'll just replace the dead part without even thinking. (At least that's assuming that you can ever get an OEM's technical support to actually do anything.)
OEMs buy in bulk. They care about quantity and about how much (how little) the parts cost to put a system that sounds good together.
If OEMs cared about quality, they'd be avoiding VIA and on-board components like the plague. They'd sell systems where each 'card' is actually a card, and where the systems never have problems.
And the day I see even 25% of all OEMs doing this, then I'll think that OEM CPUs are quality.
Honestly, why would an OEM care about high standards for the CPUs when they use the cheapest motherboards, memory sticks, and cards to ever exist? If you can give me a good answer to this question, I'll reconsider my stance.
If the opposite of pro is con, what is the opposite of productivity? Ground first.
FWIW, <A HREF="http://forumz.tomshardware.com/modules.php?name=Forums&file=faq¬found=1&code=1" target="_new"> check this out</A>.
Don't know if there is any difference. I do have a concern.
Where do OEM parts come from? I always thought OEM parts are to be used in complete systems, manufactured by big name companies and value-added resellers. So why are there so many loose parts sold to the general public?
I don't have any answers just questions.
Doesn't stop me from buying OEM. (That's excluding hard disks. I've had too many fail over the years. I prefer to have the warranty offered with retail drives).
I think a lot of small OEM businesses / resellers buy OEM parts in bulk and then offer them for sale individually, passing the savings on to you. (And making a tidy profit in the process since they're resold to you at considerably higher prices and/or higher shipping than what they cost originally.)
If the opposite of pro is con, what is the opposite of productivity? Ground first.
Wow, you guys RULE! Thanks for all the input. I was a little alarmed about the other article I read that someone linked to about OEMs vs. Retail. The posted said to stay away from the Asus A7M266 at all costs because "you only have 1 exclusive interrupt PCI slot, so what good is it to anyone?"
Could anyone shed a little light into what he means by that? What are your opinions on the Asus A7M266?
| Quote : <b>The posted said to stay away from the Asus A7M266 at all costs because "you only have 1 exclusive interrupt PCI slot, so what good is it to anyone?" |
</b>
I think what he meant is <font color=red>A7M266</font color=red> only has the PCI slot #3 not sharing IRQ with other devices, but I think he mistakened <font color=blue>A7V</font color=blue> series which have <b>Promise controller</b> on board with <font color=red>A7M266</font color=red> that doesn't.
PCI slot 1 shares with AGP.
PCI slot 2 shares with Promise controller in A7V series. I think in A7M266 this one is not sharing with other.
PCI slot 4 and 5 (or AMR slot) share with each other as well as USB controller.
In my opinion, I don't think PCI sharing is bad with modern motherboard. In my A7V133, I have my Promise controller shares IRQ5 with my Linksys 10/100TX NIC as well as my SB Live!, modem and USB controllers sharing IRQ3 without any problem.
Good or Bad have no meaning at all, depends on what your point of view is.<P ID="edit"><FONT SIZE=-1><EM>Edited by khha4113 on 06/14/01 05:20 PM.</EM></FONT></P>
Consider the difference between OEM chips that are, well, actually used by OEMs (Compaq, Micron, etc.), and OEM chips that are, well, end up on the online resellers' virtual shelves. Since we never see the former, let's talk about the latter.
The same guy who claimed retail CPUs are better than OEM (Steve Benoit from Stable Technologies, Inc.) also mentioned something about big manufacturers having a lock on certain grades of CPUs.
Leo
<P ID="edit"><FONT SIZE=-1><EM>Edited by LeoKor on 06/14/01 06:03 PM.</EM></FONT></P>
As long as people are still reading Stable's commentary, I think you should read <A HREF="http://forumz.tomshardware.com/modules.php?name=Forums&file=faq¬found=1&code=1" target="_new">this</A> and decide for yourself. This is really something that none of us can know 100% for sure. It's a gamble to some degree, but I think the odds are way in our favor for buying OEM since none of us expect to use these processors so many years from now. 3 years from now the CPU's will surely be running at 3 Ghz or way more....
<font color=red>Amd or Intel? Who cares?? Not me...</font color=red>
Well at least we both agree that the other is wrong!
Um - I think some of the disagreement stems from our definition of OEM. My definition of OEM is Compaq, Dell, HP, etc. I wasn't thinking of the 'tiny' OEM that virtually buid to order.
That said, that is why they care about quality. The Dells of the world do not install cpus that will die in 6 months time. They would be forced to recall them at their own cost. Businesses that take machines from the companies would go spastic and the OEM would likely lose contracts.
The machines that (big) business buy are all ususally 100% stable and long life.
-* This Space For Rent *-
email for application details
We're all arguing stuff that we can't prove. As far as I'm concerned, you can't go WRONG either way. So those that don't feel comfortable with OEM, that's fine, but I am as are many, so I'll save the money.
<font color=red>Amd or Intel? Who cares?? Not me...</font color=red>
why dosen't somebody just right to amd or intel and see if they can get a response and then go form there
Heh heh. Yeah. Screw it! Everyone is wrong. There is no right. There are no truths.
(Wouldn't that screw up science a lot? Heh heh.) If we can't laugh at ourselves, who can we laugh at?
Seriously though, my definition of OEM is a business that puts parts together into a whole computer and sells that computer. Both large and small companies are OEMs.
Specifically though, I would like to bring up the following company: Packard Hell ... oops, I mean Bell.
Have you SEEN some of the tricks that they have done? Including but not limited to: soldiered-on onboard memory so that you can never upgrade your memory speed, a single PCI card functioning as the modem and sound card WITHOUT jumpers to disable one or the other, onboard video WITHOUT a jumper to disable it, and everyone's favorite, a single IDE HD plug on some motherboards. (Because after all, why would anyone ever want to upgrade an OEM system?)
My mother's old Packard Hell had all of the above in a single PoS system. And worse, not a single item in that computer was from a top-notch company. (Probably because no company concerned with quality would do such things.)
Out of the list that you mentioned, HP is the ONLY one that I would trust to use all quality components.
Go look at the specs for the computers that they sell. How many use VIA motherboards? How many use onboard video? How many claim to be 'ultimate gaming machines' and yet are using cards like the Rage Fury MAXX instead of a at LEAST a GeForce2 GTS or RADEON? How many even specify if their memory sticks are CAS2? (For that matter, how many P4 systems even come with PC600 or 700 RDRAM instead of PC800?) How many are still offering things such as a 100MHz FSB 850MHz chip for MORE than a 133MHz FSB 800 MHz chip?
The day that you can answer that none of the OEMs that you have listed are doing any of these commonly used tricks to cut corners and sell crappy systems that sound better than they really are is the day that I'll believe OEMs actually care about quality.
I have seen a few OEMs in the past that don't do any tricks like this to cut the cost on putting a system together. And you know what? They were ALL small businesses, not these big brand-name OEMs. And you know what else, most of them used retail parts whenever possible instead of buying bulk OEM parts.
If the opposite of pro is con, what is the opposite of productivity? Ground first.
I don't want to leave the impression that I'm in Stable's camp on this matter. I was just pointing out a (supposedly knowledeable) contrarian side to the issue.
<b>IMHO, the decision criteria is actually a 'no-brainer':</b>
If the buyer is NOT going to overclock the CPU, then the Retail Box makes a lot of sense (and is cost-effective). You get a meaningful 3-year warranty (from the factory), and a factory-approved HSF that will do the job just fine.
If the buyer IS going to overclock the CPU, then an OEM part is much more cost-effective. Here, any warranty that you get (regardless of duration) becomes null and void as soon as you OC the CPU. Furthermore when OC'ing, the factory HSF should be replaced with a more efficient unit. So why pay the premium for the Retail Box (losing the warranty and HSF in the 'bargain')?
Hey, stable is a good guy so don't worry about that. It's just that it doesn't end up being as easy a decision as you say. First off, what is the point of a 3 year warrenty when nobody ever uses it? Second, what is the point of spending $50 more for retail at least when 2 years down the road you can just buy a better processor for those $50? It's better to replace the cpu with a NEW and faster processor when the time comes. If your motherboard can't pull that off, then you just buy the same processor hopefully for dirt cheap.
IF there was a quality difference I would have a different point of view, which I did for a while, but I just don't believe that anymore after everything I've read and seen.
Case in point: I bought a Duron 800 in December of 2000 for $85 bucks or so. I can buy one today for $33. Why in the world would I buy retail in that case? Of course buying a PII 400 today still costs over a $100, but at that point it's better to do a full upgrade if the computer dies.
<font color=red>We are going to have peace even if we have to fight for it. - Eisenhower</font color=red><P ID="edit"><FONT SIZE=-1><EM>Edited by dhlucke on 06/16/01 03:49 PM.</EM></FONT></P>
I think the main reason to go with one over the other is quality. I would expect better quality from retail than I would from OEM. That isn't to say that OEM has worse quality. Chances are, most of the time they'll probably be the same quality. I just believe that the chances of getting a part that isn't quality is higher in OEM products than retail products.
Really, when over clocking is involved, you're taking a chance either way because not all chips are guaranteed to overclock well. Sure, you can try to go with batches known for overclocking well, but if you aren't getting one specifically from that batch, then you're taking something of a risk either way. And in such cases, I personally would trust retail over OEM with such a risk.
If the opposite of pro is con, what is the opposite of productivity? Ground first.
There are 1274 identified and unidentified users. To see the list of identified users, Click here.
You are about to answer a thread that has been inactive for more than 6 months.
If you still wish to proceed, please ensure that your posting is original and does not duplicate or overlap any prior responses to this thread.

