Sign in with
Sign up | Sign in
Your question

AMD FX-6100

Last response: in CPUs
Share
a b à CPUs
October 14, 2011 1:11:18 PM

I was just wondering and I thought I'd ask it here.My question is is why is the FX-6100 so bad?

I know AMD didn't spend the last years making a product that would just be equal with the previous generation of CPU's(Phenom ii).I'm thinking either it wasn't ready or their's some sort of driver error with it.Just doesn't make sense to me why they would release a product if it's only equal to their last gen.

Any insight would be appreciated.Thank you.

More about : amd 6100

a b à CPUs
October 14, 2011 1:12:29 PM

according to what has been said the scheduler wasnt as good and the pipeline was way too long.Not sure about cpu arch still a newb at it so yeah.
m
0
l
a b à CPUs
October 14, 2011 1:22:16 PM

But it just doesn't make sense why they would release it.They tested/benchmarked long before they released it i'm sure and they would know how close it performs to the Phenom ii's.But if they knew that, why they would release it just doesn't make sense to me.

Maybe they felt they would loose money if they kept delaying it?
m
0
l
Related resources
a b à CPUs
October 14, 2011 1:40:16 PM

Desktop benches are the only thing i've seen out right now, and im guessing that this, like said even by reputable reviews, is not a desktop chip. It is a Server and Applicational chip based for heavy multitasking. Although AMD did hype this chip up for gamers, i dont think the developers ever had intentions with 1st release to have this be the "solve all" desktop chip like an i5 or i7 is from intel. Also remember, while intel has had this technology, this is AMD's first crack at integer cores (basically HT), and the first 8-core processor, true core or not.

While i am disappointed at the design for desktop use, as a server chip, this may be a huge win for AMD. It is yet to be shown yet though.

also, the reason they lose money is because it costs money to manufacture and produce testing chips. remember, they aren't just plug and play chips in the research labs, they are pushing them to absolute max, and then some, and doing it multiple times, making sure of everything this chip is capable of, and trying to improve on it. Millions if not Billions of dollars later, this is what they had to offer before what i can imageing a very ugly shareholders meeting occurred.
m
0
l
a c 123 à CPUs
a b À AMD
October 14, 2011 1:47:52 PM

In the next two weeks, I expect a "scapegoat" will be found at AMD. It will be someone who gets booted out, but possibly with a golden parachute (details of which may not be fully disclosed). This will be done to placate shareholders and save the stock price.
What will be of interest to us is who this person is. If it's an engineer, my guess is that AMD is hosed; they missed design goals. Despite possibly excellent server performance, the high TDP likely rules this out in high-density datacenter environments; the performance gain isn't worth the power (and cooling) costs. That it performs no better than existing tech leaves it out in the consumer market. The "golden" parachute may in fact be lead.
If it's a marketing droid though, there is hope; hope that there's a readily fixable bug, or perhaps more importantly there are specific, well-understood software modifications/optimizations and/or stepping changes that will let this chip shine. The guy let go "took one for the team," and his parachute may well be platinum; in two years, maturing software will have caught up to the design, and it will turn into something amazing.


Just don't expect Intel to be twiddling its thumbs during that time. "It goes on..."
m
0
l
a b à CPUs
October 14, 2011 1:52:42 PM

I just want to hear it from an Engineer themselves.They are the only ones who really knows what went wrong or what exactly they were trying to do.I'm sure no one from marketing or more higher up the knows the tech specs.
m
0
l
a b à CPUs
October 14, 2011 1:52:53 PM

not twiddling, but definately enjoying the extra time they now have in the consumer market.
m
0
l
a b à CPUs
October 14, 2011 2:26:48 PM

Do you guys think a revision or stepping will make any difference besides make the power consumption better?
m
0
l
a b à CPUs
October 14, 2011 2:31:26 PM

purple stank said:
Do you guys think a revision or stepping will make any difference besides make the power consumption better?

Remember this is a server cpu abuse and made to be a desktop.
m
0
l
a b à CPUs
October 14, 2011 2:33:25 PM

i have no experience or know how when it comes to architecture of cpus, and will never admit to having any...lol. BUT, like i said in this post, i don't believe for one minute they ever intended this first chip to be "gamer ready". I believe it to have been specifically designed for server and application use, and that within the next revision or 2, this will be a viable option for gamers. However, there are just way too many feelings on why this chip is not gaming material, and quite frankly, not enough testing from the consumer market.
m
0
l
a c 83 à CPUs
a b À AMD
October 14, 2011 4:01:29 PM

Keep in mind that new chart we see from AMD now. 5-10% per year improvements to the arch/chip over the next 4 years.

One rumor floating around is that they had to release SOMETHING and this was the design that was farthest along. I don't quite think it was THAT bad, but I suspect its closer to the mark then anyone wants to admit. We know from when AMD was talking about modules that they've been working on this for a long time. And because they aren't as big as Intel they likely weren't working on much else. They've been tweaking the basic K8/K10 arch for a long time so I'm not surprised to see them needing more years to bring BD up to those specs.
m
0
l
a c 123 à CPUs
a b À AMD
October 14, 2011 11:08:52 PM

purple stank said:
Do you guys think a revision or stepping will make any difference besides make the power consumption better?

If there's a L1 cache bug, absolutely; could be 10% right there. If other efficiencies altogether added another 10% (include BIOS revisions in there too), we could see a 20%-25% improvement, which pretty much puts this chip where it needed to be; if not a thriller, certainly no disappointment.
m
0
l
a b à CPUs
October 14, 2011 11:28:30 PM

Everyone's seems to be topping out at around 4.7-4.8ghz on air.Do you think if the revision/stepping fixed the power consumption "issues" that you could get higher clocks? If their is a next stepping/revision what do you guys think would be fixed?

Do you happen to know the general area for which Sandy Bridge processor's top out at on air? I was hearing 5ghz but i'm not sure.
m
0
l
a c 83 à CPUs
a b À AMD
October 14, 2011 11:58:34 PM

For air yes. 5GHz or so for SB is "normal". I've heard others talk about being able to cross 5GHz.
m
0
l
a b à CPUs
October 14, 2011 11:59:37 PM

That was their reply so far, doesn't make any sense though !!
http://blogs.amd.com/play/2011/10/13/our-take-on-amd-fx...
it's said that Win 8 will take more advantage of the 8 cores and thus BD might work better. I'm also waiting for more reviews on diff boards, new rev or new BIOS.
According to Wikipedia, there will be a 2nd generation of BD codenamed "Piledriver" which will improve the performance by 10%
and a 3rd generation codenamed "Steamroller", 4th codenamed "Excavator"
the "Piledriver" is to be released 2012 and i guess they will be working on it so hard in order to fix their situation
m
0
l
a b à CPUs
October 15, 2011 12:10:07 AM

I would assume with the rest of the Bulldozer lineup coming somewhere in the Q1 2012 that piledriver won't be out until maybe Q4 2012.

Is piledriver gonna be like what they did with the AM2+ Phenom ii 's to AM3 Phenom ii's?
m
0
l
a b à CPUs
October 15, 2011 12:14:06 AM

no details and no one can make expectations, but for sure they will prepare for a massive surprise or they will be vanished soon.
and we are waiting for their official response on the current release of FX
m
0
l
a b à CPUs
October 15, 2011 12:16:13 AM

purple stank said:
I would assume with the rest of the Bulldozer lineup coming somewhere in the Q1 2012 that piledriver won't be out until maybe Q4 2012.

Is piledriver gonna be like what they did with the AM2+ Phenom ii 's to AM3 Phenom ii's?


possibly but the real issue is, AMD stream lined their R&D process for server and PC CPUs. most likely a cost cutting move to please stock holders so one CPU can serve two masters. as usual with jack of all trade products/people, they are also masters of none.
m
0
l
a b à CPUs
October 15, 2011 12:50:50 AM

I found it a little strange how stealthy they were about launching Bulldozer.No advertising or official release dates prior to launch.Which leads me to believe that they weren't happy with what they made.Seems like they might have been rushed as someone stated in an earlier post.
m
0
l
a b à CPUs
October 15, 2011 12:52:39 AM

be noted that BD was under production 2-3 years ago, it was originally made to compete with the first I7 gen, so i guess it's in the wrong time.
m
0
l
a c 82 à CPUs
a b À AMD
October 15, 2011 1:35:17 AM

http://www.tomshardware.com/reviews/fx-8150-zambezi-bul... here is bulldozers problem. It is clear and present, per core performance is bad and that is why is is bad overall. I dont know what AMD was thinking, expecially considering its worse than a phenom II core. Logically, as myself and a few other people have said before, they would be better off modifying the phenom design and cutting their losses with bulldozer. An 8 core phenom II made at a smaller process would be faster period.
m
0
l
a c 80 à CPUs
a b À AMD
October 15, 2011 3:39:03 AM

amd's vista (dozer) to be followed by 7 (piledriver, 7sp1 - trinity) and 8 (excavator)? :p 
m
0
l
a c 123 à CPUs
a b À AMD
October 15, 2011 4:30:44 AM

Yeah, or like Phenom followed by Phenom II, with a few stepping improvements thrown in for good measure. I think that's a best-case scenario though, so I'm not planning to hold my breath. My next full build may not be for years, and although I'd anticipated a CPU upgrade, it isn't going to be THIS Bulldozer.
m
0
l
a c 123 à CPUs
a b À AMD
October 16, 2011 10:46:05 PM

Well, software fixes may help a measurable amount, but it really looks like there's some hardware issue too, or at least that it simply isn't optimized for the tasks that gamers (and/or the rest of us) run. I really don't think we'll have a clear idea for another few weeks, maybe more.
m
0
l
a b à CPUs
October 16, 2011 11:03:43 PM

i'm very optimistic about the Piledriver, and i guess also that AMD might make new chipsets or change the existing AM3+ socket
m
0
l
a b à CPUs
October 16, 2011 11:12:22 PM

Well since Bulldozer is geared more towards servers maybe we can hope that Piledriver is more geared towards gamers and enthusaists.
m
0
l
a b à CPUs
October 17, 2011 12:48:57 AM

what's strange is that there's no BIOS update for existing 890FX boards, only a few released for high end AM3 Asus Mobos and no one reviewed compatibility with it
m
0
l
a b à CPUs
October 17, 2011 12:51:25 AM

BIOS update to accept AM3+ CPU's? or performance improvements?
m
0
l
a b à CPUs
October 17, 2011 1:00:07 AM

Yes, BIOS update for the new Zambezi to work in existing AM3 boards, for eg My MOBO is revision 2.0, revision 3.0 and 3.1 were produced with the new AM3+ socket but with the same 890FX chipset, even though there's no BIOS update for the revision 2.0 and 2.1

It's said that the new Am3+ socket is 942 Pin and AM3 is 938 and FX 8150 is 938 pin, that's really mysterious
m
0
l
a b à CPUs
October 17, 2011 1:21:02 AM

So maybe it's just a matter of time till i get my BIOS update ? but according to your MOBO the BIOS update is available since 2011-04-02 ? that's weird too, no one got to confirm backward compatibility on that date
m
0
l
a b à CPUs
October 17, 2011 1:38:22 AM

My mobo was weird from the start.MSI accidently put AM3+ on it in lettering.It's above the northbridge.It looks like your mobo is discontinued so I'm not sure if they will still provide support for it.
m
0
l
a b à CPUs
October 17, 2011 1:46:04 AM

They provided BIOS update for the AM2+ MOBOs after the AM3 ones appeared. Something is missing..., still the question why AM3+ is 942 pin ?
m
0
l
a b à CPUs
October 17, 2011 1:50:48 AM

i remember reading a while back it had something to with the voltage or something.It's a small difference but it would require a new socket.
m
0
l
a b à CPUs
October 17, 2011 1:59:42 AM

i read a lot about the new AM3+ architecture and it wasn't mentioned about the 942 Pins even Wikipedia has useless info about it. anyway we still need to know how FX performs on the AM3 MOBOs, is it giving the full potential power of it or not.....

And i guess i'm not going to update right now, maybe wait for 2012 CPUs
i was really thinking to upgrade to a BE CPU cuz i have problems overclocking the 1055T
m
0
l
a c 82 à CPUs
a b À AMD
October 17, 2011 2:17:45 AM

^problems OC'ing your 1055?? you have it at 4.1ghz, that is an insanely good overclock on air.
m
0
l
a b à CPUs
October 17, 2011 2:19:42 AM

That was before the BIOS update, i'm only able to get a 3.8 and not stable sometimes... something goes wrong.
Flashed back to stable one, no success also.
m
0
l
a c 83 à CPUs
a b À AMD
October 17, 2011 4:09:18 AM

purple stank said:
So I was doing some browsing and I happened to stumble upon this.

http://quinetiam.com/?p=2356

Do you guys think their's any truth in that?


Be amazing if true, but I doubt we'll see that kind of performance gain. Something interesting that mentions is they lost performance to remain backward compatible, that could very well be true. It be interesting to see what would have happened with an all new socket design for the new architecture, since its still technically AM3 compatible.

Either way, I'll be buying another Phenom II X6, or an Intel 2500k/2600K before any of the current Bulldozer based processors.
m
0
l
a c 83 à CPUs
a b À AMD
October 17, 2011 4:54:49 AM

Some of the places ran tests on win8 which was supposed to understand BD better. Many tests scored better, but I think on the order of 10%, not 40%. There is a lot more performance to be had from the chip, but I seriously doubt they can do much about the single core performance problem.
m
0
l
a c 80 à CPUs
a b À AMD
October 17, 2011 5:41:24 AM

purple stank said:
So I was doing some browsing and I happened to stumble upon this.

http://quinetiam.com/?p=2356

Do you guys think their's any truth in that?



some things i noticed:
the guy sure rants a lot, calls people names.
imo they're right about losing performance for backward compatibility, amd's attempt to build a cpu for intense multi-threaded work, server capability, trying to blur the definition of 'cores', need of software optimization, trying to innovate and increase performance at the same time.
they don't mention the full configuration especially what role the gpu played or if quick sync was in action.
he does admit that 'it's all about trinity, komodo and terramar'. then what's left for bd?
tries to underplay bd's lack of pcie 3 support as 'not needed'. i thought video editing and rendering needed a lot of data bandwitch and pcie 3.0 would help with that.
they needed liquid cooling to reach 5 ghz+. i observed from different bd reviews that it barely goes over 4.4 ghz on air cooling , 4.7 if you're lucky and 4.9 on liquid cooling. i guess oc has a lot to do with luck.
he says windows 7 is not fully optimized for bulldozer.. iirc sandy bridge came out after win 7 did.. i don't see those cpus or phenom iis underperforming.i guess re-benching with updated drivers and patches will show some improvements for bd.
and what about those really really high power consumption figures with and without overclocking?
i am no video editor or hardware expert. i just want to use my pc for browsing, watching videos and may be play some games, mess around with some home video clips and photos once in a while. bd seems to be less power efficient for the task i'd need it to perform.
i may be ranting like that guy but i wrote this only because he incorrectly accuses sites like tom's and anandtech for 'thrashing bulldozer and shutting down..' etc. etc. very uncool.
there is no absolutely objective review. reviewers simply share their experiences with the products they're reviewing.
m
0
l
a b à CPUs
October 17, 2011 7:35:42 AM

O.C.ing does have a lot to do with luck.Each chip is different.

Maybe Bulldozer is more like a break the ice type of thing so they can learn how to harness the best out of this new arch.
m
0
l
a b à CPUs
October 17, 2011 9:18:00 AM

purple stank said:
I was just wondering and I thought I'd ask it here.My question is is why is the FX-6100 so bad?

I know AMD didn't spend the last years making a product that would just be equal with the previous generation of CPU's(Phenom ii).I'm thinking either it wasn't ready or their's some sort of driver error with it.Just doesn't make sense to me why they would release a product if it's only equal to their last gen.

Any insight would be appreciated.Thank you.


The main problem I think is its viewed as a 6 core chip. I don't view bulldozer cores as true cores in the contemporary sense. The module is comparable to cores as we have come to think of them. The second integer core in the module is just AMD's version of hyperthreading. But "worlds first 8 core" was just too good for AMD marketing to pass up.

When you examine the 6100 from that perspective, its a tri core chip and suddenly your looking at a significant per core improvement over Phenom. Unfortunately it didn't improve on Phenoms gaming performance which is a big deal in the enthusiast community. Thus the outrage.

Another issue, and another poster touched on this, is load scheduling. Bulldozer threads are like hyperthreading. The loads need to be scheduled one per module before loading the second core. This is because the cores on the same module share resources and therefore only run at full speed when the second core is idle. Windows does not yet do this with Bulldozer modules. This can kill performance in lightlly threaded apps and that appears to be happening with bulldozer. This part of why the gaming performance disappointed. This issue is an OS issue and is not going to be fixed until windows 8.

m
0
l
a b à CPUs
October 17, 2011 9:25:35 AM

iam2thecrowe said:
http://www.tomshardware.com/reviews/fx-8150-zambezi-bul... here is bulldozers problem. It is clear and present, per core performance is bad and that is why is is bad overall. I dont know what AMD was thinking, expecially considering its worse than a phenom II core. Logically, as myself and a few other people have said before, they would be better off modifying the phenom design and cutting their losses with bulldozer. An 8 core phenom II made at a smaller process would be faster period.


Actually, there have be tests of the modules with the second core disabled which bypasses the aforementioned OS scheduling issue by simply disabling the second core on the module and the IPC improvement is considerable.

The IPC improvement, roughly 20%, moves Bulldozers IPC ahead of Phenom and is a preview of how much performance is being lost to the scheduler.

m
0
l
a b à CPUs
October 17, 2011 9:42:35 AM

Onus said:
In the next two weeks, I expect a "scapegoat" will be found at AMD. It will be someone who gets booted out, but possibly with a golden parachute (details of which may not be fully disclosed). This will be done to placate shareholders and save the stock price.
What will be of interest to us is who this person is. If it's an engineer, my guess is that AMD is hosed; they missed design goals. Despite possibly excellent server performance, the high TDP likely rules this out in high-density datacenter environments; the performance gain isn't worth the power (and cooling) costs. That it performs no better than existing tech leaves it out in the consumer market. The "golden" parachute may in fact be lead.
If it's a marketing droid though, there is hope; hope that there's a readily fixable bug, or perhaps more importantly there are specific, well-understood software modifications/optimizations and/or stepping changes that will let this chip shine. The guy let go "took one for the team," and his parachute may well be platinum; in two years, maturing software will have caught up to the design, and it will turn into something amazing.


Just don't expect Intel to be twiddling its thumbs during that time. "It goes on..."


You forget ACP.

That AMD regulated spec which some people compare with TDP.
m
0
l
a b à CPUs
October 17, 2011 6:30:29 PM

FALC0N said:
Actually, there have be tests of the modules with the second core disabled which bypasses the aforementioned OS scheduling issue by simply disabling the second core on the module and the IPC improvement is considerable.

The IPC improvement, roughly 20%, moves Bulldozers IPC ahead of Phenom and is a preview of how much performance is being lost to the scheduler.


Do you have any benchmarks to prove that?
m
0
l
a c 83 à CPUs
a b À AMD
October 17, 2011 7:43:09 PM

I would have added "from reputable sites" to that comment Purple. On launch day I saw some benchies from Win8 which was supposed to be smarter about BD, but it still wasn't faster overall then a 2500k.
m
0
l
a b à CPUs
October 17, 2011 8:27:07 PM

purple stank said:
Do you have any benchmarks to prove that?

Good lord people. This stuff has been all over these threads the last few days. Read of few pages before you post.

http://translate.google.com/translate?hl=en&sl=auto&tl=...

There you go. 20%, well actually 19%. And they are not the only ones who have done this. Most have been hitting plus 10% to 20% or so with a few as high as 30%. Depends on how its done. For example, Behardware came up with 11% but included multithreaded apps which are not expected to benifit from scheduling.

If bulldozer had been 15% faster in lightly threaded apps, the reviews wouldn't have been nearly as harsh. It isn't going to suddenly beat the 2600k, but its a nice boost and should open up a clear lead on the X6 1100 and better position it against the 2500K.

m
0
l
October 17, 2011 10:58:28 PM

purple stank said:
Do you have any benchmarks to prove that?

One has to accept this redefinition of AMD's Modules to Cores, to come up with this kind of IPC improvement.
m
0
l
a c 83 à CPUs
a b À AMD
October 17, 2011 11:46:21 PM

See purple, thats why I added that bit that I did.

Great if true, but I prefer not having to use google translate in order to read your proof. If "This stuff has been all over these threads the last few days." I would think it would be easy to find links. I've seen sites that show improvements with Win8, so I know its very possible.
m
0
l
!