AMD ''Vishera'' FX-Series CPU Pricing Leaked

[davemaster84]

with Piledriver 'Vishera'
single core performance will ONLY be as good as Deneb C3 hopefully better but not close to Intel.[/quotemsg]

I don't know where did you get that information, but since actually there are no real benchmarks around to prove you wrong all I can say is that the AMD's expectation followed by a lot of users and reviewers is that Vishera will have at least 10% over thuban (best phenom around) and around 15% over bulldozer.

 

[proffet]

Thuban is not better than Deneb C3 (per core).

the scaling (from 4 to 6) of the Thuban six core shows it to be slightly less of a hitter than Deneb C3 quad core.
look it up.

 

[nebun]

[citation][nom]luciferano[/nom]A well-configured FX-8120 can meet or marginally beat those i7s in performance with only marginally higher power consumption at about half the price. Also, their called third-generation i7s.[/citation]
if you call 77W marginal when compared to AMD's 125W offering there is something wrong with your logic

 

[mousseng]

if you call 77W marginal when compared to AMD's 125W offering there is something wrong with your logic

If you think an FX with half its cores disabled still has a 125W TDP there is something wrong with your logic.

When he says "well-configured" he means disabling every other core and overclocking it, as blazorthon and others have said before in other places. I should also assume he is referring to gaming performance when he means that it can beat an i7.

 

[nebun]

If you think an FX with half its cores disabled still has a 125W TDP there is something wrong with your logic.

When he says "well-configured" he means disabling every other core and overclocking it, as blazorthon and others have said before in other places. I should also assume he is referring to gaming performance when he means that it can beat an i7.

then why purchase a multi core cpu if you are going to disable the half the cores?....you must be very smart, much smarter than most of the people out there

 

[army_ant7]

Don't the pricings and specs look a little unusual? I mean 16MB cache for the 8 cores and 8MB for the 4 cores, but wouldn't you expect 12 instead of 14MB with the 6 cores?

Also, notice the $10.95 priced difference between the FX 8350 and 8320. Who wouldn't pay for that amount of money for a 500MHz difference if you plan to use it at stock and the chances of having a higher binned chip for overclocking?

Well, even if these may or may not be retail prices/MSRP's, we still oughtta take these with uncertainty along with the specs, wouldn't anyone agree?

 

[army_ant7]

then why purchase a multi core cpu if you are going to disable the half the cores?....you must be very smart, much smarter than most of the people out there

Because if that processor ends up having a better price/performance ratio with a core per module (not just any 4 cores)... Don't talk like that. We could use less close-minded thoughts here.

if you call 77W marginal when compared to AMD's 125W offering there is something wrong with your logic

I just want to point out that if you think TDP ratings equate to power consumption numbers, then your logic may not be flawed, but your knowledge/information may be.

You could be more pleasant with your words.

There was another method shown to possibly be better than disabling a core per module. It's a method using PSCheck to basically manipulate Windows thread scheduler to the way you'd want it to work, and you get to keep all 8 cores, though it involves changing their clockspeeds so I'm not sure how highly-threaded performance comes out with this compared to no modification. Just learned it from blazorthon, which he in turn learned from palladin(some numbers)...

 

[mousseng]

then why purchase a multi core cpu if you are going to disable the half the cores?....you must be very smart, much smarter than most of the people out there

To be clear, we're talking on the grounds of gaming performance. That said:

- An i5 is a quad-core, and roughly $200 on average.
- An FX-8120 is an octo-core, and typically about $160.
- If, by disabling half of the cores on the FX (which, for gaming, will not be used regardless) you can achieve levels of performance nearing that of the i5 (at both max performance and performance-per-clock comparisons), why spend the extra ~$40?

Don't the pricings and specs look a little unusual? I mean 16MB cache for the 8 cores and 8MB for the 4 cores, but wouldn't you expect 12 instead of 14MB with the 6 cores?

Surely it's a typo; each enabled module is allotted 4MB of cache (as it is with Bulldozer), so I don't see how it even could have 14MB of cache (unless AMD is doing something silly with Piledriver).

 

[halcyon]

To be clear, we're talking on the grounds of gaming performance. That said:

- An i5 is a quad-core, and roughly $200 on average.
- An FX-8120 is an octo-core, and typically about $160.
- If, by disabling half of the cores on the FX (which, for gaming, will not be used regardless) you can achieve levels of performance nearing that of the i5 (at both max performance and performance-per-clock comparisons), why spend the extra ~$40?


Surely it's a typo; each enabled module is allotted 4MB of cache (as it is with Bulldozer), so I don't see how it even could have 14MB of cache (unless AMD is doing something silly with Piledriver).

I don't mean to ask the seemingly obvious question but why get an AMD processor with the goal of meeting Intel performance when you can just get an Intel processor. Is there a shortage of Intel processors that we were not told about?

 

[proffet]

I don't mean to ask the seemingly obvious question but why get an AMD processor with the goal of meeting Intel performance when you can just get an Intel processor. Is there a shortage of Intel processors that we were not told about?

+1000, I second that.
plus isn't the 2500K/3570K in the same pricing of the 8150 at Microcenter and other such places.?
on-line retail might be different.
but the 'in-general' $40 difference is worth it to me because of all the other determining factors including
the previous mentioned power consumption.

If you just do not like Intel then say so but to say that AMD (with all the tricks like disabling half of it) is almost an Intel equivalent is asinine.
maybe in one or two benches but that's it.

so many dweebs looking like AMD fanbois..

 

[halcyon]

+1000, I second that.
plus isn't the 2500K/3570K in the same pricing of the 8150 at Microcenter and other such places.?
on-line retail might be different.
but the 'in-general' $40 difference is worth it to me because of all the other determining factors including
the previous mentioned power consumption.

If you just do not like Intel then say so but to say that AMD (with all the tricks like disabling half of it) is almost an Intel equivalent is asinine.
maybe in one or two benches but that's it.

so many dweebs looking like AMD fanbois..

I understand trying to back AMD but AMD has to participate as well. They're choosing to let Intel bitch-slap them and I don't like a quitter...sorry.

 

[rdc85]

Can I put an request here ?

When there a benchmark of these babies, could u include pII.. It might old and add more workload for the review..

But there still lot of people out there that still hold on pII (me included), The info will had much value...

 

[army_ant7]

The problem is mostly caused by a lack of x86 decoders (four per module compared to three per core in *K10* and four per core in every high-end Intel arch since Core 2 AFAIK) in each module, so I think that my analogy is accurate. The houses could be said to be conjoined, but the processing isn't the issue which is what I think that your analogy would imply, the decoders feeding the cores are insufficient, so the cores simply can't be *fed* as AMD would put it quickly enough. Consider this:
Two people have 8 hot dogs total to eat. This is not a problem when they can both use one hand each to eat, but if they had to eat with only one hand between both of them, although they'd probably eat faster than one core, they'd have a bottle-neck. The hot dogs would all be able to get to the hand quickly enough, but the hand simply can't do the work of two hands and it can't feed both persons as quickly as two hands could although it can feed one person without any issues at all.

Steamroller intends to fix this by giving each core four decoders instead of sharing four decoders between each core of a module.

That two guys, one hand (That just sounds wrong! :lol, and a bunch of hotdogs (That even sounds wrong...er!) analogy was pretty funny, though when I visualized it, it did seem valid as to that fetch-decode problem. Good head!

I don't understand this. How can they put CPU and GPU in the same package with decent results and how can they NOT design a well performing CPU in a dedicated package?

Because maybe those decent results are based on the GPU's performance and not just the CPU's (two different things)?
Also, Vishera is the "full" CPU implementation that's supposed to perform better in CPU tasks (depending on the task, like how many threads it can tax well). Along with more modules, there's also the L3 cache.

I don't mean to ask the seemingly obvious question but why get an AMD processor with the goal of meeting Intel performance when you can just get an Intel processor. Is there a shortage of Intel processors that we were not told about?

+1000, I second that.
plus isn't the 2500K/3570K in the same pricing of the 8150 at Microcenter and other such places.?
on-line retail might be different.
but the 'in-general' $40 difference is worth it to me because of all the other determining factors including
the previous mentioned power consumption.

If you just do not like Intel then say so but to say that AMD (with all the tricks like disabling half of it) is almost an Intel equivalent is asinine.
maybe in one or two benches but that's it.

so many dweebs looking like AMD fanbois..

No. No shortage. Just the fact that you can save money and end possibly end up with the same results, as was previously mentioned. Also, power consumption could actually be better when modified if what luciferano says is true (I'm not sure if it was on this thread but it was less than the i5's power consumption.), again, as was mentioned. Not to mention the possibility to compete with the i7's in heavily-threaded workloads. Whether you think the extra $40 is worth it your own personal preference, maybe you don't take pleasure in tweaking or maybe you're just too lazy. I understand that until you see numbers, you'd be skeptical. But dispelling the possibility is foolish IMO. Calling the ability to reach another more expensive product asinine seems ridiculous to me, well, maybe you're one of those "money is a none issue" kind of people.

And like blazorthon usually says, not everyone has the privilege of having local stores with those prices.

 

[proffet]

all you guys blinded by listening to 'blazorthon' as he trolls the articles and threads are all headed down the incorrect path....
I for one will not fall into the trap of subliminal marketing and fanboism for AMD by way of a thread troll...
that's asinine.

my last comment on this matter, peace out...

 

[halcyon]

all you guys blinded by listening to 'blazorthon' as he trolls the articles and threads are all headed down the incorrect path....
I for one will not fall into the trap of subliminal marketing and fanboism for AMD by way of a thread troll...
that's asinine.

my last comment on this matter, peace out...

You surely must have misunderstood Blazorthon...I've never seen him troll. He typically works hard to remain neutral from my perspective.

Now me, on the other hand. Oh...ohkay...guilty as charged.

 

[mousseng]

I don't mean to ask the seemingly obvious question but why get an AMD processor with the goal of meeting Intel performance when you can just get an Intel processor.

Because some folks would rather put that $40 somewhere else - say, a better CPU cooler or video card or hard drive. I'm not saying everybody should do this because it's a better way, I'm saying some people should consider this because it might be a better choice for their budget.

plus isn't the 2500K/3570K in the same pricing of the 8150 at Microcenter and other such places.?

Not everyone has access to Microcenter's prices. In fact, I'd say a majority of people don't.

but the 'in-general' $40 difference is worth it to me because of all the other determining factors including the previous mentioned power consumption.

As has been mentioned (several times) before, power consumption shouldn't be too different between a quad-core FX and an i5.

If you just do not like Intel then say so but to say that AMD (with all the tricks like disabling half of it) is almost an Intel equivalent is asinine.
maybe in one or two benches but that's it.

I don't like either one more than the other, each have their place in the market. If you want benchmarks for something like this, then check this out to see the improvement BD getsper thread when each core doesn't share resources (compare the 2CU/4C and 4CU/4C results). These numbers give much more generous 30%+ improvements, but we can stick to a more conservative 20% if you like (which is what I generally quote anyway). Here's a comparison between an 8120 and a 2500k - you can scroll past all the synthetics and go straight to the games since that's what this kind of tweak is for. When you adjust the FX's score by a factor of 1.2, a lot of games would look to run as well as the i5, if not better (keep in mind, some of them look to be bottlenecked by the OP's 6950).

so many dweebs looking like AMD fanbois..

Because we're suggesting something you don't agree with you resort to insulting us? Very mature.

 

[halcyon]

Because some folks would rather put that $40 somewhere else - say, a better CPU cooler or video card or hard drive. I'm not saying everybody should do this because it's a better way, I'm saying some people should consider this because it might be a better choice for their budget.

Not to be a booger, but if $40 makes that much difference on something as important as the CPU, I'd say it should just stay in your bank until you can afford the more adequate product. ...just one perspective.

 

[mousseng]

Not to be a booger, but if $40 makes that much difference on something as important as the CPU, I'd say it should just stay in your bank until you can afford the more adequate product. ...just one perspective.

This hinges on whether or not the "more adequate product" actually is more adequate - or at least $40 more adequate. For some, it may seem to be, and for others it might not. And for the record, I do appreciate your perspective.

 

[army_ant7]

all you guys blinded by listening to 'blazorthon' as he trolls the articles and threads are all headed down the incorrect path....
I for one will not fall into the trap of subliminal marketing and fanboism for AMD by way of a thread troll...
that's asinine.

my last comment on this matter, peace out...

blaz, does everyone a favor by sharing what he knows/has heard of instead of making negative, useless remarks like you do. We are not blinded by him or following down the wrong path. We are just presenting other possibilities/options that aren't as much in plain sight.

You don't have to believe it, but if you have no proof against it, why give us your close-minded crap? You're being "asinine" as you like saying. If your comments are just gonna be like this then it's for the best that that is your last comment on the matter. Come back when you can be useful or at least more pleasant with your opinions.

 

[halcyon]

This hinges on whether or not the "more adequate product" actually is more adequate - or at least $40 more adequate. For some, it may seem to be, and for others it might not. And for the record, I do appreciate your perspective.

You're reading my frustration with AMD's giving up in the performance-cpu market. I must admit, it bothers me.

 

[army_ant7]

You're reading my frustration with AMD's giving up in the performance-cpu market. I must admit, it bothers me.

Don't worry, you're not alone. I don't think even Intel fans who analyze this would be happy about it because Intel would have to deal with anti-trust issues if their main competition goes belly-up, also the fact that less competition tends to mean higher prices (for Intel CPU's).

We could hope they're just saying that (bluffing) so that Intel would start getting lax then suddenly they'd pull a fast one. :lol: Haha!

 

[halcyon]

Don't worry, you're not alone. I don't think even Intel fans who analyze this would be happy about it because Intel would have to deal with anti-trust issues if their main competition goes belly-up, also the fact that less competition tends to mean higher prices (for Intel CPU's).

We could hope they're just saying that (bluffing) so that Intel would start getting lax then suddenly they'd pull a fast one. :lol: Haha!

We can hope and wish.

 

[zeratul600]

it seems like 2013 its gonna be a great year! haswell and maybe i pray to god that amd comes to the surface instead of keep drowning themselves! i want a powerful plataform and i want it to be amd intel rocks but amd makes me feel like a rebel outcast! my 3770k it damn powerful but was way too expensive i hope to get a new tower for under 500 $ that can be upgraded to dual- triple crossfire eventually

 

[luciferano]

[citation][nom]De5_roy[/nom]from what i've read so far, ... intel couldn't (or intentionally didn't, to sell ivb) validate for pcie 3.0 in time for sb-e launch.intel will never sell an unlocked, <32nm dual core/triple core/core-unlockable cpu. i think they ended with i5 655k. they would do this to protect the unlocked i5 and some lower priced i7s. hell, they even make seperate cpu dies. it isn't about consumer convenience. it is, and always has been, about money. they segment their products in a way that customers have to pay more to get better desired performance. the lack of competition in the i5/i7 price segment doesn't help either... i read somewhere that because of the new 22nm design, newer 3d finfet tech etc somehow made the cpu more fragile, may be the solder would put more pressure on the cpu. another reason is the one more well-known: to sell ivb without hurting the sb sales while focusing on mobile sector. the ivb cpus have higher tjmax so stock isn't usually the problem. oc is where the heat and temp issues start and intel intentionally sidestepped them with ivb.intel is in a place where they can enforce certain things on users as no one's there to challenge them. and they are not diong anything illegal this time e.g. in ivb oc case, intel openly says that they do not cover component overclocking in their standard warranty (neither does amd).amd isn't anywhere near where intel is and they don't seem to be chasing intel that way. their biggest asset has been their way of out-strategizing intel like they did by designing the brazos and llano apus. i am expecting them to do the same with trinity and piledriver although it doesn't seem like they've succeeded so far. i really hope they turn around soon.[/citation]

I wasn't asking as if I thought that Intel made mistakes with some of those, I was giving examples of Intel's *abuse* of enthusiasts. Also, using proper paste, although not as good as solder, would have still helped greatly on Ivy Bridge, I wasn't referring to unlocked multipliers with the i3s (all i5s and i7s can be overclocked regardless of the multiplier unlocking by about 20-30% through the Turbo settings and for the more adventurous, BLCK overclocking), and I realize why Intel does things. I just don't agree with a lot of it. I also don't agree with a lot of what AMD does, but at least at this time, they're not trying to screw their customers AFAIK; they're just screwing themselves.