amd's big blunder

acejj26

Senior member
Dec 15, 1999
886
0
0
Okay, with the slot Athlon, AMD basically invited everyone (including remarkers) to overclock their chips, because I believe they thought that someone buying, remarking, and selling an Athlon was better than someone buying and then selling an Intel processor. At least AMD makes SOME profit off of their Athlon by doing so. And, in the process, they established a following of hardware enthusiasts who loved to tweak every last ounce of raw processing power out of their new Athlon.

Now with the socketed Athlon, AMD has decided to shun the following that they created with their slotted Athlon. No longer could people tweak their processor to their hearts content. Now, if you buy an 800 MHz Athlon, you get an 800 MHz Athlon. That's it. You can't try to get 850 or 900 (outside of playing with the FSB which is pretty much futile with the current chipsets and motherboards available).

What would have motivated AMD to do this?? Why would AMD want to alienate their new following and pretty much ask them to buy an Intel processor/815 motherboard?? Why would they pull a 180 and want people to buy an Intel chip that they can reasonably overclock via FSB modification than buy an AMD processor?? I thought that the perfect system to buy now would be a socketed Duron, overclock it like mad, then down the road pop it into a 760 based board with some DDR memory. Guess maybe that P3/815 doesn't look so bad now.

Also, what happens when motherboards based on the 760 chipset come out?? This will feature a 133 MHZ DDR bus. Now they should still be using socket A, so a socket duron/t-bird should work in that motherboard...but it probably won't. Why?? Well if the multiplier is burned into the processor, we'll use a 6x for a duron for example, the 600 MHz Duron we buy now will have to run at 800 MHz in a 133 MHz system, no?? If we can't drop the multiplier down to 4.5x so that it still runs at its rated 600 MHz, we have a useless chip, which means NO UPGRADE PATH!!! The KX133 fiasco was bad enough, this will be worse because it will have happened twice in a row. Does AMD not see this?? Hey, I guess they figure we'll have to buy all new chips and motherboards just like they're hoping we do now with KT133/Duron or TBird. More money for them, right?? Maybe not...some of us Athlon adopters may grow a bit weary of AMD's actions to alienate all of us and switch platforms.

I guess this is AMD's new marketing strategy. Gain a following with a kick-*ss product that people will love, execute really well with that product, then slowly begin to alienate all those who bought it until they're back in the position they were in before they released the product.

I hope that AMD realizes this fatal flaw. They should remember that if a lot of Athlons are remarked, that means they still sold a lot of Athlons, even if they are losing some revenue based on the fact that remarkers are selling lower speed processors as higher speed ones and taking away some of AMD's high-end sales. They're still sales, nonetheless, right?? Now I don't condone remarking at all, but I think AMD has to look at the big picture. They need those of us who look at the Athlon as a great chip, whose main advantage over the P3 is the ability to play with the multiplier. If I have a choice between a 700 MHz Athlon (which should be able to hit 800 to 900 MHz using standard bus speeds) or a 700 MHz P3 (which can hit 800 MHz using non-standard bus speeds), I take the Athlon. Now if you give me a choice between a 700 MHz Athlon which can MAYBE hit 750 using non-standard bus speeds or a 700 MHz P3 which can still hit 800+ using non-standard bus speeds (and utilizing a better chipset, nonetheless), I may think twice about buying the Athlon.

I hope AMD sees this and rethinks their strategy. Remarkers take away some potential profit, but it also shows that a lot of people like to buy the Athlon as a great overclocking processor, which is good for AMD. If AMD thinks that they will be boosting profits by locking their processors, I think they're in for a rude awakening. They will see their following, which they have built up over the last year, begin to dwindle. I have been loyal to AMD for a while since they have always been friendly to the overclocking community (even though the K6 was an awful overclocker, they still left the option open), but that loyalty is beginning to fade. By locking their processors, they no longer have that one edge they had over the P3.

Please share your thoughts about this in hopes that someone from AMD will read this and they will rethink their strategy. They should know what we all think about them locking their processors, potentially providing no upgrade path to the 760 chipset, and potentially losing the following they created over the last year.
 

Czar

Lifer
Oct 9, 1999
28,510
0
0
Nicely put, send it to AMD and post their response here to ;)

What sucks is that now AMD disables overclocking and Intel goes Rambus, whats left to do then think about Cyrix *shudder*
 

KarsinTheHutt

Golden Member
Jun 28, 2000
1,687
0
0
AMD doesn't give a damn about hobbiests like u and me. The % of people who actually overclock and tweak their systems is probably 1% or less.

AMD will need to convince big buisiness to buy systems based on its chips if it wants to do well in the long run - I suppose they don't want an unscrupulus OEM fiddling with the chip and then selling it in a system out of spec.

AMD can also afford to pull the rug out from under us because they have the most bang for the buck right now. If P!!!'s were more competetively priced, you can bet AMD wouldn't be locking those multipliers...

I woudn't worry too much
 

KarsinTheHutt

Golden Member
Jun 28, 2000
1,687
0
0
I heard those Cyrix III chips overclock pretty well... I think Tomshardware got a 500 MHz one to 800+ MHz. The only problem is the FPU is so bad that it would have to run at 1.6 GHz to catch up with Mendocino Celeron!
 

Train

Lifer
Jun 22, 2000
13,586
82
91
www.bing.com
ya thats what happens when a company grows and becomes successful, the marketers take over and mess everything up for everyone, Intel did it, now AMD is doing it, we will have to wait until the next new player in the game, and rely on them to make great products for a few years, then only to get taken over by more marketing retards.
 

Xanathar

Golden Member
Oct 14, 1999
1,435
0
0
I really dont see your point. As far as I can see so far <I may be misled> but the only multiplier burned in is that of the MAX multiplier. I see this as the smartest move they could make because it stops remakring of the chips, but still allows us overclockers flexability. What happens if your duron 600 cant hit 800 at 133 bus? then you just drop the multiplier by half and you can hit 733. Or better yet, what if you memory can handle 150mhz ddr, but your chip cant handle the speed, you just drop the multiplier a tad.

Wouldnt that be designing for the future? realize current motherboards and solutions are generally the amd 750 which was designed not even to be a production chipset, and the kx133 which was VIAs first experience with the EV6 Bus protocol.
 

pm

Elite Member Mobile Devices
Jan 25, 2000
7,419
22
81
The part that you and most of the users on this forum don't seem to understand is that remarking is not performed by a few minor thieves in third world countries. It it not a minor issue involving a million or two dollars of lost revenue. It is a huge business typically run by mafia-like organizations in dozens of countries. Not only does the company lose revenue on a massive scale, but it also loses its reputation for quality.

The remarkers don't really care if the processor is reliable at the higher speed grade... it just needs to appear to work. So a customer buys one of these processors and it ends up being unreliable - who do they blame? The processor is very professionally marked as a 800MHz processor - and yet it doesn't run at 800MHz. Who's fault is that? Of course it's the manufacturer's fault. So they send the processor back to the manufacturer, all the while thinking &quot;This company doesn't build reliable products&quot;. Then the company sends another processor, and wastes engineering effort trying to figure out what is going wrong.

The thing is that overclocking is a very minor effect practiced by an insignificant minority of the computer user population. In recent years, the press has started to cover overclocking and it has moved beyond the &quot;tech junkie&quot; crowd, but still the vast majority of computer users don't know what overclocking is, don't know how to do it, and probably would be scared to try even if they did know. On the other hand, remarking is a widescale problem that results in an indirect and direct loss of revenue, in prestige, and in consumer confidence. It turns a manufacturer into a detective and law enforcement officer and takes away effort from the company's primary goal - to build and sell high-quality processors.

Patrick Mahoney
IA64 Microprocessor Engineer
Intel Corp.

* Not speaking on behalf of Intel corp. *
 

acejj26

Senior member
Dec 15, 1999
886
0
0
xanathar,

Yeah, according to Anand, you can drop the multiplier by using pretty sophisticated techniques. That does leave the upgrade path open, to a degree. But the main point of the rant was the fact that a lot of us would like to INCREASE AND/OR DECREASE the multiplier at will.

By doing that, we won't have the Intel debacle with having to release different chips that run at the same speed (500E, 500EB, for example). By letting us increase or decrease the multiplier as we see fit, they can sell a 600 MHz chip which we can run at what ever speed we can get it to run at and whatever multiplier and bus speed we want. If Intel did this, they wouldn't need to have a 'E' or 'EB' differentiation; when someone wanted to migrate to a 133 MHz based sytem, they could use the same chip but with a different multiplier that they choose. This saves on advertising, marketing, documentation, and packaging. AMD is going down the same path and once the 133/266 based systems come out, if the multiplier is locked, AMD will have to differentiate between a 600 MHz chip running at a 6x multiplier and a 600 MHz chip running at 4.5x multiplier.

Also, it just sucks knowing that AMD has been dangling that &quot;carrot&quot; of the ability to change the processor's multiplier at will &quot;right in front of us&quot; only now to take away that &quot;carrot&quot; altogether. So, now, what makes their processors so unique from Intel's. Cost?? Well, couple a $50 cheaper processor with a $40 more expensive motherboard, and the difference isn't that great anymore. Plus, just by using the fsb, the Intel processors are much better overclockers. So now AMD can't differentiate the Athlon from the P3 by performance or the cost difference, which when you look at the whole system, is becoming negligible. AMD NEEDS to keep the ability to change the multiplier (either up or down) to have that one leg up on the P3. They've been able to sell Athlons for the past year because they've been overclockable. Take that away, and who knows what will happen...
 

Czar

Lifer
Oct 9, 1999
28,510
0
0
What AMD and Intel doesnt see is that us overclockers and hobbiest are often those who work at computer shops or administrate big networks. Its us that choose what is for sale in shops and it is us who choose what to buy for the big networks. Actually its not that much &quot;us&quot; but its the field we are biggest in. And one other thing, the people who dont know much about computers come to us and ask us what is best to buy, I´m sure that for the last 6months most of you have been recomending AMD systems just because we know they are better for the money. Its so much about goodwill.
 

KarsinTheHutt

Golden Member
Jun 28, 2000
1,687
0
0
I dunno Czar, corporations and gov't still won't even consider AMD. And I've seen tons of salespeople push the P!!! systems even though they are more expenseive and perform about the same.
 

acejj26

Senior member
Dec 15, 1999
886
0
0
I understand that remarking may reflect poorly on the manufacturer, in this case AMD. But the fact remains that AMD built their first Athlon just asking for people to overclock them. &quot;Hey, let's put a little connector on the cartridge so people can adjust the multiplier and the voltage. Let's give them the ability to solder resistors to change the multiplier, voltage, and cache settings.&quot;

Is that not begging for people to remark their processors? AMD did this so that they could build a following of loyal customers who buy their processors knowing fully well that they were going to overclock them (and in turn, remarkers then did the same modifications and resold them). Now they turn their backs on the AMD faithful who either can't afford or don't have the expertise to play with the &quot;golden bridges&quot; yet they still leave the option available to remarkers who have the capital to remark them en masse.

I'm just a little peeved that AMD is doing this because for the last year, AMD has sold their processors with the ability to change the multiplier as we see fit, and now they are ditching that policy. My point is that since I'm peeved, I'm sure there are a lot of others who are as well, and a whole lot of people turning their backs on AMD will not bode well with them. My feelings are that the money they will lose on those of us abandoning the Athlon platform for Intel will be greater than selling their processors to someone who remarks them, yet still counts as revenue for AMD.
 

Rogue

Banned
Jan 28, 2000
5,774
0
0
If I'm not mistaken, the FSB and the memory bus are two separate entities, unaffected by the other. I don't see how a clock multiplier built into a processor is going to prevent it from working properly on the AMD 760 DDR chipset. Am I wrong on this? I mean, right now I can run PC133 memory at 133MHz on my Epox EP-7KXA board but still be using a 100MHz DDR=200MHz FSB. All the while, it doesn't matter which clock multiplier I use and it leaves the other two (memory and FSB) alone completely. I think someone is wrong in his assumptions...
 

Czar

Lifer
Oct 9, 1999
28,510
0
0
Like why doesnt AMD put something in the BIOS so that the BIOS tells the user that you are running a 600mhz Duron but overclocked to 800mhz.. so the user knows that it is actually a 600mhz Duron. It would end all remarking and make us very happy.. nothing more pleasurable as seeing your friends see a 550mhz Athlon at 850mhz :p
 

acejj26

Senior member
Dec 15, 1999
886
0
0
Rogue,

The FSB I'm talking about is when the 100 MHz DDR bus becomes a 133 MHz DDR bus, which is what the CPU uses to talk to the chipset. By raising the bus, you effectively raise the processor speed as well. Right now, the asynchronous bus you are talking about on your system is the memory bus and that's it. You are still running the processor on a 100 MHz bus. Once that is bumped up to 133/266, your 750 MHz Athlon (for example) becomes 1000 MHz, and if you don't have an easy way of changing that multiplier (via BIOS or dipswitches), you either can't upgrade to the better chipset or you better hope like hell your Athlon can hit 1000 reliably, because if it doesn't, your chip becomes worthless.
 

NFS4

No Lifer
Oct 9, 1999
72,636
47
91


<< If I'm not mistaken, the FSB and the memory bus are two separate entities, unaffected by the other. >>


In VIA Apollo Pro 133A chipsets and the KT133/KX133/KM133 that may be true (you can set the memory bus independent of the FSB).

But on Intel chipsets, the memory bus and FSB are one in the same to my knowledge.
 

acejj26

Senior member
Dec 15, 1999
886
0
0
Czar,

BINGO!!!!! That was what my idea was. I'm sure that wouldn't be too hard to implement, except AMD didn't think about that when designing the processor and they didn't put any microcode onto the die to let the system know what the rated speed was. If they did that (ala Intel), then:

They don't need the &quot;golden bridges&quot;;

Remarkers can't get away with telling someone they are getting a &quot;true-blue&quot; chip when they aren't;

I wouldn't have had to write this rant, and everyone would be happy.
 

Rogue

Banned
Jan 28, 2000
5,774
0
0
I do know what a FSB is and I understand the concept, I just can't see that a higher FSB will somehow overclock the processor in the manner that you state. I firmly belive that AMD will implement some sort of 2/3 multiplier or something to enable the processor to run within spec. Don't count your chickens before they hatch.
 

Adul

Elite Member
Oct 9, 1999
32,999
44
91
danny.tangtam.com
Why be mad at AMD for making a business decision? If you should be mad at anoyone it should be the people who remark the processors. They are the ones who ruined it for us. While I may not like AMD's decsion on the multiplier, we are still a small part of the market. Look at it this way, would you sell a business a system that is overclocked and running stable? What if that cpu happen to burn out? They would be stuck buying another CPU to replace one that voided its warrantee.
 

BD2003

Lifer
Oct 9, 1999
16,815
1
81
First I hear you cant O/C durons, then I read all these articles about the Asus and Abit MB's that you CAN change the multiplier with and oc'ing 650s to 900, and now this....is anyone else confused? Whats the deal?
 

Rogue

Banned
Jan 28, 2000
5,774
0
0
I'm with you BigDee2003. What the hell is going on here? Tom has found a way around it according to him and Anand is saying it can't be done. What gives?
 

Adul

Elite Member
Oct 9, 1999
32,999
44
91
danny.tangtam.com
I still see a problem with putting microcod on the CPU to tell whats it speed is rated at. even if the microcode was changable, what keeps remarkers from make changes to that microcode has well. Burn it in you say. Ok, now AMD can not change its own microcode on a processor that it may need to sell at a lower speed to help fill in a demand.
 

jpprod

Platinum Member
Nov 18, 1999
2,373
0
0
AMD-760 will surely support 100mhz DDR FSB (200mhz) settting as well, what's this talk about not having a platform upgrade path?
 

KarsinTheHutt

Golden Member
Jun 28, 2000
1,687
0
0
I like Czar's Idea.

Chip makers could have the processor's rated speed included in the processor like the P!!!'s ID number (heh heh heh). If its etched into the silicon, I don't think remarkers can change it without damaging the chip, or at least not without expensive equipment.
 

Noriaki

Lifer
Jun 3, 2000
13,640
1
71
Ya BigDee2003, I agreee with that.

I've been sitting here thinking.
Hmmm Display the Duron's proper clcok speed, and then allow you to change the multi. setting. Isn't that exactly what the BIOSes in the Asus A7V and ABit KT7 do?

I'm certianly confused...
 

acejj26

Senior member
Dec 15, 1999
886
0
0
Okay, if the 760 supports 100/200 DDR, you then can't take advantage of the faster 266 MHz memory. The chipset basically becomes the 750 with DDR memory support. I would just like to know that when I upgrade, I can use the newer FSB with my same chip until I decide to upgrade my chip.

I know this was a business decision. The main point, then, is why did AMD decide to make their first wave of Athlons so overclocking friendly, which in turn paved the way for remarkers so that they then take away our ability to overclock. It's a logic flaw on AMD's part - Let them overclock the processor very easily, and then take that ability away when they do overclock it. Does anyone see the logic in that? I don't.

And for the record, yeah, I'm p*ssed off at the remarkers for forcing AMD to make this awful decision. Because of them, that new Intel setup is gonna look pretty good in a few months once the prices come down. The 815 chipset will offer better AGP and memory implementation than anything VIA has out/will have out. Pretty soon, I am going to build a computer for one of my friends, and that socket A TBird and A7v was looking really nice until Tom Pabst got that email from AMD saying that all production socket A Athlon/Durons will be multiplier-locked. Now, maybe I go with a 600 -> 800 P3 on an 815 board for his computer. I'm still not sure yet.