Results From the Latest Hammer Poll

We had a poll the other day.

Page 1 of the poll painted a scenario in which AMD kept relatively high prices on its socket 939 and future 90nm chips, while selling Parises fairly cheaply. Readers were then asked what they do given the scenario, and how happy they’d be about the situation.

The initial scenario also asked people to assume that AMD would change its mind and make some top-end socket 754s rather than relegate socket 754 to the value-processor Paris. It then asked those who said they’d want top-end socket 754s what they’d do if that option weren’t available.

Here’s what you said:

1) What do you do?

2) If these end up being your choices, how would feel about them?

Analysis

There seems to be three distinct groups among AMD fans out there:

The Happy Campers About a third of those who responded seem happy enough with the choices and prices AMD is likely to give them, and will buy in due time.

The Fat and Happy Generally, these people have late-model socket As (fairly often, very recently topped up socket A systems), and they’ll be perfectly happy to wait until they get an offer they can’t refuse, either by price and/or new/additional features. This constitutes a bit less than half the population.

Just Unhappy About 20% just don’t like what they see. They have all kinds of reasons for not liking what they see, from pricing to muddled roadmaps to inadequate increases in performance, but whatever it is, they don’t like it.

What was prevalent among all the respondents was a decided lack of enthusiasm for Hammers. Even the happy campers pretty much just indicated they’d buy one, without any display of excitement.

That’s not a bad thing, of course, that’s the way people should be when they decide to spend their money. You want sales, not seizures. But if AMD thinks people are foaming at the mouth waiting to buy these things, that is not what is happening in the trenches. The only foaming at the mouth that is starting to occur out here (and not even too much of that) is not due to enthusiasm for AMD, but rage against it.

Only about 3% of those who responded indicated in some way that they already owned a Hammer system, so this audience has a track record of controlling their enthusiasm.

Socket 754

There is very little interest in Paris at the moment, though I suspect it is still over the horizon for many.

There was decidedly more interest in socket 754, provided AMD kept it up to date.

When we asked those who wanted a 90nm socket 754 what they would do if AMD didn’t give them one, they said:

2) If these end up being your choices, how would feel about them?

The purpose behind the AMD plan to make socket 754 the budget processor is to keep it from cannibalizing socket 939 sales. While these answers indicated this is true to some extent, these numbers point to the possibility that AMD would lose more in net sales than they would gain in socket 939 sales.

The Best of Your Comments

These generally drift from positive to negative:

So prices are somewhat higher. I don’t think it’s unreasonable to pay a little more for something good. Prices definitely can’t drop too much
from what they are, otherwise, they degrade and corrode their lower end chips. I think AMD deserves a few bucks more for
their nice work and effort. I’d definitely keep them in my mind when purchasing new systems for my friends.

I have had thoughts of getting a 754 system and it has been hard to resist more so with the word from limited amounts of 939
processors. I am itching to upgrade my Skt A system. As of now, I don’t care about DDR2 due to very high latency. Also getting sick
of hearing delays.

I already own an Athlon64 3200+ so I won’t upgrade to socket 939. I’ll wait till dual core hammers come out. My next major upgrade
will be a dual core system. At this point I’m happy with the performance of my A64 system and feel no need to upgrade. I had a
duelie before and miss the multitasking performance and would like a dual core chip.

I’ve already got an athlon 64 system here. I’m looking forward to a 90nm chip, preferably in a more reasonable price
point on s939, but if not I can deal with continuing to use s754. I have water cooling, so heat
isn’t an issue to me. It’s the fact that this chip was brought out on a process
with the end so close in sight, which left very little OC room, not to
mention the “issues” with the various chipsets. 90nm, on
either socket should help to alleiviate those problems.

It’s hard to say what I will do. My current position is a 2500+Mobile running at over 2.8ghz on phase change with a relatively high fsb for an Athlon XP. Being at the top end of cooling, and with some pocket change always laying around, I always have the urge to jump on board when new chips come out, though I always try to think logically about my choices and expenditures. Currently, my XP Mobile is (it’s hard to say this being a speed junkie) more than enough for me. Anything more right now I feel would just tend to add to my false sense of Badass, but generally no one cares what is in your comp but yourself and a few friends so I am inclined to wait until there is a much more significant edge and price point. Knowing that I will have to replace my mobo, ram, and vid card also adds to my apprehension to jumping on board. Though if I do get anxious, socket 754 looks like a cheapo alternative for 6-9 months or so before 90nm gets established and we see what’s up with dual cores and the memory prices stabilize in the DDR2 market. As I assume I should be able to keep my current ram/video setup on some 754 boards. This way I just dip my foot in the pool instead of jumping head first and then crying when the water is too cold.

I don’t have any immediate plans to upgrade my socket A system. If something comes out that I just HAVE to have…maybe. None of the assumptions you listed are something that I HAVE to have. Pretty much the only thing that’s going to make me upgrade at this point is a “killer app,” most likely a game that I want to play that I have problems running on my current system. And probably then…only because I’ll need a PCI Express platform to get a decent video card.

Right now I have 2 dual socket A systems both running around the 2000+ (1.8GHz) range. I am probably not typical, but I don’t think it would make much sense for me to upgrade in 04-05 if those are my choices. I might go from my current Tbred chips to some faster mobile Bartons and get more RAM, but I don’t think I’ll get enough advantage going to Hammer until (if it ever) x-86/64 becomes mainstream. It would be nice for some reasonably large advances to be made, so I can play with upgrades and overclocking more, but I don’t see it happening too fast. Also, my current software is running great, and though I haven’t played DOOM 3 yet, I can’t think of anything that I can’t run smoothly on my system. To me this says save my money for later.

Frankly, I have an Athlon XP at 3200+. I really dont need any more
until the killer app comes along that requires more. The only apps that really require all that CPU are video
manipulation. Thus, I have one system to render and compress for
hours on end and the other to surf and play with. So, one might
consider that I already have more CPU cycles than a 4000+ would
afford me, and I bought them (combined) cheaper than your proposed
3300+ option. I spent about $250 on computer guts last year ( $40
AN35N MB, $80 XP Barton to overclock, a stick of PC3200 memory and a
$100 9600pro), and I expect to spend even less this year ($0 so far)
and next.

When the new technologies are all together with the bugs ironed out
(SATA drives, PCI Express devices,new case footprints and power
supplies,new motherboard and chipsets for the new processors and
cases) THEN maybe it will be worth it to me to upgrade. I have to
wait for both the introduction of these technologies and the
debugging process to produce stable stuff. Then, I want an OS and
some applications that take advantage of them. This really all starts
to sound very post-2005 to me.

I would not buy a 939 system until the Board+CPU cost was around
under about $325. Thanks a lot for the heads up on the Athlon XP not going 90nm, my
dreams have been shattered and now I need a cup of coffee so I can
save up for my 939 (1.5 years). If prices look like they’re going to
drop considerably at the dawn of 2006, I would put off the overhaul.

Feeling this constant need to upgrade at a cost of thousands of dollars isn’t a choice I’m going to be forced to make and I don’t doubt there are many others who feel the same way. If they want to sell me new products then they better find a way to make it quieter, smaller, or for a similar good reason. While in some ways it might look cool to have an antec server tower on my desk it really feels like more of a necessity to have something so large to maintain cooling. If I were given the option of buying a new S939 90nm system that ran stock on passive cooling with a 2000+ PR rating then I’d take a much more optimistic look at buying their new toys. Silence is golden. But no way I’m upgrading for a 15% gain in performance over an Athlon XP system for around $2000 CND. Period.

The whole thing is a stupid mess, I just hope mobile socket A will last me till cheap 90nm 939. New graphics cards are $499, so a proc and mobo being $300 isn’t too bad

I will switch to Intel at this point not only due to price but due to a slightly more stable upgrade path. I’m certainly not saying that Intel’s future is brighter or clearer but the majority of purchasers are Intel and any mistakes by Intel are “more” likely to be supported/worked around somehow or someway. With such a huge installed base it makes more sense to put your money with the majority especially when the minority is playing games.

I like AMD for the competition they bring to the table but I do not see a future with the current choices the management is making. they are quickly losing my vote. It seems to me that they need to simplify the lineup of the future. Too many close choices available may lead to really bad purchases. I’d hate to see AMD discontinue one of their paths after they find out its not profitable and leave users stranded. This “all things to all people” approach will end up biting them in the ass.

If Intel has 64 bit, and it’s maybe $100 more, then I’ll at least know if I have upgrade paths. I want DDR-2 with a wide enough bus to use it, pci-express, and 64 bit. I want the board, chip, and memory for no more than $400 combined, and I want the video card to be better than today’s geforce FX, and about $100. If I can’t get that, then I’ll wait. I’ve been frustrated about there being nothing going on in the computer world since Thoroughbred B. Honestly, the only upgrades I’m interested in now are a gigabit network, a 140gb raid array of 10K rpm hard drives, or an LCD monitor, and the latter two are still too expensive for me to bite.

Unless the NX bit turns out to be a killer app I don’t need 64, but I want an on-die memory controller baby! Athlon XP is just fine otherwise. And I already have 2 XP’s at 2.2+GHz/400fsb on quiet air that cost $47 (Tbred in 2002) and a $70 barton. Will buy at least 1 more XP before it’s too late.

I’m going to wait until Intel releases desktop Pentium-M chips,
dual-core or otherwise. Then see what AMD’s response is. I am
becoming more concerned with a quiet and cool computer than a speed
demon. Speed demons are no longer the only motivating factor in
buying a PC. My mobile Barton 2.4Ghz will last long enough until I see what AMD
*and* Intel have for their next generation parts in 2005.

My loyalty is to my dollar, not my CPU company. I’ll wait until AMD’s products return to the glory of the Thunderbird era; faster than Intel, heavily overclockable, and cheap too. “Cheap” is defined as <= $65 for the chip. And I want an unlocked multiplier. Yep, I might have a long wait in front of me, but I don't game hardcore like I used to, and my Athlon XP is rock stable at 2400 MHz, which is more than enough for D2 and Alpha Centauri. If I'm very lucky, my Mobile Barton and NF-7S mobo will last through another year of high school and four of college.

[My last upgrade] got me a 40% improvement in real system throughput, not just
another miniscule 2-4% that nobody can notice. I don’t want to
upgrade again until I can get at least another 40-50%. I don’t
really care primarily about cost – what I care about is constantly
“fiddling” to get tiny increments in performance.

Let me give you my impressions as well. The ONE thing I see as
impacting true performance of a computer is gaming. (The only other impact is
internet speeds, but that is a different topic.) I have an o/c
Barton 2500+ in dual memory mode with RAID (stripped) and when I capture videos of
my daughter from my analog camcorder, there are no lost frames. When I
rip MP3s and WMAs, it is fast. When I work with MS Office, it flies. So
what hangs me up? Games.

What is my next purchase for my NF7-S? An ATI 9800 PRO or an ATI
X800 XT. That will keep my rig flying for years – yes, years. In 2006 or 2007,
I might have to drop down to 600 x 800 and turn the settings to
low/medium, but in reality, unless all software goes 64-bit, I have no reason to
upgrade the MB+CPU. Uless someone comes out with a hulking 64-bit gaming
engine, why do it?

I’ll never pay $400.00 US for a processor again. If I were going to upgrade it would be a socket 939 system, however I will not pay more than $200.00 for a processor. The last time I paid $400.00 for a processor was when I purchased an AMDK62-400 back in 1997. At that time $400.00 processors were pretty cheap for top end. Now a days, computers are so fast that they perform most tasks just fine and there is no software that I run that really requires an upgrade from my current system (KT7A-RAID and XP 2100+ running at 2200Mhz actual speed and 768MB of PC133 RAM). I don’t play games, and mainly use the system for video editing, digital photography, surfing the net, e-mail, and word processing. Having more power to process the video would be nice, but it’s not worth the money to upgrade if I have to lay out $400.00 just for the processor. By the time you add up the processor, motherboard, video card, etc you’re looking at $800.00+. Too rich for my blood when my current system is meeting my needs fairly well. If I have to encode a large amount of video, I just kick the job off before I go to bed and it’s complete before I get up in the morning.

The AMD picture is too muddled for me – I’ll wait until things settle down.

It will be a while before Windows software to take advantage of 64 is widely available, and the 32 bit difference is not worth the $. Linux is a different story but I don’t use it that much. I’ll stick with socket A until I see significant price/performance gains, and some substantial overclocking fun.

Socket 754 is a very reluctant choice.

In the near future (2005) I will not pay more than $150 for a CPU,
period. I can afford to wait, my “3600+” is not obsolete
quite yet. If AMD will only offer me a $150 CPU in S754, I am going to be very
picky about my spending.

If I do not feel comfortable with any of my S754 choices, and S939 is
not an option, I may find a new hobby for a while.

If AMD introduced an S939 Paris, or, in general, any <$150 CPU for S939, they would be much more successful in parting me from my money. I approve of x86-64 in general, but am not willing to pay extra for it - until most of my applications (games) support it.

I understand the need to increase the pricing of their processors (You can’t stay in business on less then a zero profit margin) but trying to make more profits by doubling the price of the latest processor line is crazy. The current customer base will drift over to the Intel camp and there is no way you can draw in more customers from the Intel side by jacking up their process to match Intel’s. The only thing that will happen from this current pricing scheme is make the AMD loyal customers PAY to be loyal. Personally I am not that loyal. I love the articles you write and AMD should read and heed the words.

To tell you the truth I don’t think I will be buying a
desktop ever again, well at least not what we call a desktop
today. Maybe some sort of porta PC or tablet PC something of
that nature, but not a desktop. ALl this bull**** that has
been happening with Intel and AMD made me re-think my
computing needs and I decided that notebook is the way to go
at least for the next 12 months.
I recently purchased an alienware notebook (loaded) and can’t
be happier it is powerfull enough to run just about anything,
including current and future games. Insides are fully
upgradable (mem, HHD, CD-R, video, LCD screen etc.) except
motherboard and as an extra its portable, h-h-mmm 10~lbs!

I personally don’t feel that desktops are superior any more,
except for maybe the ability to plug in the latest video card
and the number of expansion bays they offer.
I don’t see the need for anyone to have to upgrade their PC
every year. Better hold off for something worth while or at
least for something that makes sense for the future. *Latest*
doesn’t always mean the *Greatest*. And unfortunately neither
Intel nor AMD are making it easy for us. I believe that a
laptop is the way to go this year and most of the next.

If the total package falls outside of my ability to pay, I simply
cannot afford to upgrade. It’s possible that price premiums associated with
PCI Express, DDR2, 64bit CPU’s, possibly having to buy a new case to
accommodate a new motherboard form factor and increased power consumption which
means a new PSU. I also believe that Video Cards are heading up as was
indicated recently when mid range cards entered the market at $200 and dropped
to below $150 (my price point for video cards) at the end of the chipset
lifecycle…. Aka the days of the Ti4200 are gone.

It would be one thing if AMD increased their price by $50-$75 but I
believe everyone will be jumping on the band wagon. Because of this I doubt
that I would be able to upgrade at all in 2005, which is sad for me.

Conclusion

Not exactly blood-curdling rabid fanatic enthusiasm, is it? If this is how the AMD fans feel about it, what about the non-fans?

BTW, if there’s a bias to the comments, it’s in the direction of enthusiasm, not away from it. I tried to include all the most positive comments, but you can’t put in what isn’t there.

Please note that I’m not trying to imply that there’s any kind of blood-curdling rabid fanatic enthusiasm on the Blue team. That doesn’t occur at the best of times, and these are hardly the best of times for Intel.

What is important to note, though, is that AMD is in danger of believing its own PR. Outside of a small proportion of AMDroids, AMD fans are not x86 versions of Mac fans; they won’t coming running just because Uncle Heckie says so.

This is a group that looks at the goods and the price, not the hype. These are not going to be gimme sales from a captive audience. Frankly, those polled seemed more in the market for a new video card than a CPU.

What AMD needs to understand is that for this group, it’s not a choice between AMD or Intel. For this group, it’s a choice between expensive AMD vs. cheap AMD, a Hammer versus a mobile Barton, and so far, the mobile Barton is wiping the floor with the Hammer.

AMD should not overestimate the value of Hammer. It’s core buying audience sure isn’t.

P.S. Thanks to all who participated, and thanks for the very nice comments about the site many made!

Ed