Separate names with a comma.
Discussion in 'AMD Processors' started by Mr. Bluntman, Sep 24, 2015.
The 40% number is not new information. Although the idea that it can be better than 40% is and this time they mentioned what the 40% was based on. I say we will have to wait some time to see how much the clock speed decreases versus bulldozer to see where this stacks up versus Intel.
Supposedly that 40% is against the unreleased Excavator. So, maybe 45-55% better than what we're seeing now? That would put it in Ivy Bridge or Haswell territory. It would definitely close the gap a bit, but AMD said Phenom, then Bulldozer would return them to being competitive. Color me skeptical, but I hope for every enthusiast's sake they pull it off.
It would be awesome to rock an AMD based primary PC again.
All AMD needed to do in order to be at least a relevant big-core option was 25% IPC gains while doubling their thread count, which it looks like they will do since talk is that Zen will ship with 16 core options and an SMT architecture (multiple threads per core? etc).
If they actually pull off 40% and the price is right, Intel really needs to worry. Pricing will be everything. Pricing would only drop out of the equation completely if they were actually FASTER than Intel at stock, which is plainly impossible.
Power consumption is an issue for some people. If you get near Intel performance but need 150+W to do it, then people will still end up with Intel.
Personally, I'm hoping they pull a rabbit out of their hats because I'd love to go back to AMD.
Even in the best case for AMD, Intel need not worry.
Intel has being doing nothing on the CPU front beyond FPU improvements for certain large customers and process improvements to save themselves money on each chip produced.
The good thing is that they haven't dropped their prices, which gives AMD an ability to enter into higher priced segments. They would be STUPID to try and compete on prices. They have to offer a slightly better deal, but try and keep Intel from reacting at the same time.
That said, I do expect them to offer cheaper alternatives to the six and 8-core Intel options.
An 8-core Zen-based CPU with > Sandy Bridge IPC? Yes, please!
Just hope the clock rates are comparable!
Yeah, that's kinda what I mean. Everyone knows that you don't really need Haswell levels of IPC for gaming. If you come in and offer something like the above, and it's maybe got a great GCN 2.0 GPU on the die with VCE to handle streaming while your main GPU handles the game (something APU+GPU setups can already do now), then you have yourself a winner, so long as it's priced right below the i5 series, as the FX series currently is. If they could do that I think Intel really should worry, as it would mean losing a big number of enthusiast gamers.
My biggest fear, other than that the IPC claims are spurious, is that the pricing is gonna be like the Fury - not really cheaper enough to justify the purchase out of the gate. Even as Fury improves with driver age, and it becomes the better card than the 900 series from nvidia, people will have already passed it by, and consumers rarely give old product a second glance. To be truly competitive, AMD needs to get in the $200-210 range right off the bat, and go down from there over time to something like 180-190 a year after release.
It would close the gap quite a bit but still leave Intel ahead. That would be fine so long as AMD prices Zen competitively. No doubt they will. Of course they'll need to update their motherboard platform which is another matter. Even when their CPUs were better, AMD always lacked a bit on this front.
As for the last part of your statement, I don't get it. I couldn't care less who makes my hardware so long as it performs well and is reliable.
I was just saying after a fashion that it would be "fun" for me to have a high performing non-Intel system is all, just as I consider old PowerPC, Motorola 680x0, and SPARC systems "fun" as well. It's a novelty I get amusement from, nothing more. Don't get me wrong, I've loved the Intel systems I've had since '08, even my old iAP5T30 i430TX based Socket7 system is still going strong 18 years later - they have been rock solid as far as stability goes. You can't knock them for that, for sure.
Intel 1 to 3% new cpu
anyhow waiting for zen myself to change systems.
I expect this to be for the 4 core / 8 threaded processor. The 8 core hopefully will be at least competitive with Intel's 6 core so I expect it to be a $500+ part. Although the 95W TDP rumor has me expecting a 3GHz stock clock on the 8 core / 16 threaded part.
This is only because AMD's current cores were designed to have a lower IPC than Phenom II. They took away 1/3 of the integer ALUs as part of the process of taking 125% of the transistors of a Phenom II core to make a 2 core bulldozer module. I expect a single zen core to have more transistors than a bulldozer module and at least as many ALUs.
That sounds nice. I'll believe it when I see it though.
I can't wait and hope AMD comes through this time. I would prefer to continue using AMD purely for sentimental reasons but am getting to the point where I can't wait any longer. I agree with the article tho that if Zen is a flop then that should be about the last nail in the coffin for AMD. At least in terms of their CPU business.
Intel created the current gap with the Core 2 Duo and widened it with Nahalem. Everything between and after those CPUs has been more or less incremental. So Intel's huge gap increases came YEARS ago. AMD making a 40% leap now still won't net them parity with Intel on a pure performance front. More importantly do not forget that AMD has made some pretty audacious claims months prior to a CPU release only to go silent right before launch. Phenom was slower per clock than the Athlon X2. Phenom II was better, but didn't close the gap between AMD and Intel. Bulldozer was supposedly some 50% faster than the Core i7 950 and Phenom II x6 1100T. It wasn't even remotely close to achieving that goal.
Well you might expect that but we really don't know. We don't know if this CPU has low clocks and high IPC, or high clocks and low IPC. We really have no idea how this performance is achieved. I do expect that it will have fewer cores and better align with Intel on that front, but even that is just a guess.
Exactly. AMD has a history of making claims prior to CPU launches and being laughably far off the mark come launch day.
Well if Zen is a flop AMD still has K12 to try and bring them back from the brink of death. That's supposedly a ARMv8-A (AArch64) CPU which will be pretty close to useless for the consumer market at present. However ARM processors are doing well in the server market. AMD will certainly have to shift gears at that point as their consumer CPU business will essentially be done. If AMD sells off it's graphics division they'll get a nice influx of cash but it will only delay the Grim Reaper a short time. AMD has a history of selling off profitable divisions and divesting them in order to keep in the CPU race. A race they've floundered in since the Core 2 Duo launched.
The 40% higher IPC part is speculation.
After all this years I've learned to distrust AMD figures, not for what they say but what they don't say.
The article states the new cores are 40% faster at the same clock rate. But the new cores do multithreading so suddenly it doesn't sound that good.
IF there's a 40% higher IPC, Zen with its 16 threads would absolutely destroy anything intel has today in multithreaded applications and I just don't see that happening.
I do believe Zen can be 40% faster or more with less power draw than Steamroller and that would put it somewhere between SB and Hasswell.
They never said the 8 core / 16 threaded procesor would have a 4.7 GHz stock clock.
I think your spot on. This is all speculative, but I think your probably pretty close to the mark.
Well they didn't say much about clocks, so who knows.
I say this is nearly impossible if the 95W TDP rumor is true. GF will not produce a 14nm process that is this much better than Intel.
I think Zen is going to put Intel out of business for good.
Firstly, it's speculative.
Secondly, even if they meet the 40% figure, they will still be a good deal behind Intel.
If they meet it though, at least they will be back in the running for our hard earned cash. Right now, they aren't even a viable alternative for most enthusiasts.
Based on what?
They are saying that their chips will be "up to" 40% faster.
1.) It's an internal figure. These are always questionable, especially from AMD as of late.
2.) "up to"
3.) It would take more than 40% to catch up to where Intel is today, and this is more than a year off, at which point Intel will have advanced even further.
I'm thinking it will likely be a huge leap, but that they will still have A LOT of work to do if they want to be competitive with Intel.
I am hopeful AMD can come out with something that saves the company, gives us a competitive CPU market again, and provides something I actually want to put in my rig again (I would LOVE to go back to AMD) but this information does not look THAT promising to me.
Obvious troll is obvious.
I think that was more poking fun at everyone that has said Intel's product "X" or NVidia's product "y" or AMD's own product "z" will put AMD out of business.
Bulldozer was 50% faster than the i7 950, in exactly 1 benchmark
Jim Keller was involved, so I believe the 40%+ gain in IPC, I also believe that the end clk speed will be competitive with mainstream intel processors. Intel's performance offerings will still be on top, of that there is no doubt.
AMD's marketing folks have been an embarrassment to AMD and the industry, Zen will put them back on the map as long as the price is reasonable when released.
I'm not sure if they will release a 16 core + hyperthreading desktop CPU. Maybe they will, with some special liquid cooler but I'm going to bet it's going to be a 160 Watt CPU.
The normal desktop processors will probably be under 95 Watts
I expect the desktop processors to be the Fanboy, OEM and Hobbiest focus.... which makes me think they are focusing on advertising for the initial release.
The mobile Zen processor is the real game changer. I fully expect genuine quad core laptops to exist under $1000 due to Zen and Intel will have to make some major price changes after mobile Zen's release.
AMD doesn't need to compete, they need to provide a "viable alternative" first.
But can't deny if they have a 8Threaded chip that performs like a 3770k at a 6600k price, it will get a lot of good reception.
Specially in gaming.
Or 8c/16t for the price of an Intel 6c/12t part with motherboards being half the cost or less without skimping on features and being competitive in performance. That's a win in my book. The insane prices on boards is one of the biggest gotchas to LGA2011v3 today.
If the motherboards are as fully-featured as their competitor's platforms, then pricing can be consistent between them and it would still be a win.
That is exactly what I am interested in. I am looking for an upgrade from an i7 970. However I will not sacrifice single threaded performance. The upgrade must have both improved single threaded performance and improved multithreaded performance.
I'm as big of a Intel/Nvidia fan boy as it gets. I have NEVER owned a AMD cpu and my last radeon was 10 years ago... but if the Zen can do what it claims, I'll hop back on the AMD wagon just to support them.
The latest generation from Intel (Skylake) has been complete shite. With a complete monopoly who knows how many years of unbearable mediocrity to expect from Intel.
I say silicon has hit a wall where shrinking the node does not allow for higher clocks (than the 4GHz we have had since 32nm) at the same power. Although I do admit without serious competition at 4GHz Intel has shifted to tackle ARM at 2GHz.
If AMD can do the following, I will buy one.
- Single threaded performance on par with, or better than my 2011 i7-3930k overclocked to overclocked.
- At least 6 real cores. (I don't care about logical cores. Take them or leave them)
- No less than 40 PCIe lanes. (preferably more)
I want to support AMD, and I will buy this product, even if it has marginal or no performance gains. I will - however - not buy anything in 2016 or 2017 (whenever it comes out) that performs worse than what I bought in 2011.
I'd say its a combination of both, and one more thing that hasn't been mentioned.
1.) Successive node shrinks more and more difficult as we are getting closer to the limits of silicon
2.) No real competition from AMD means that they have not had to push as hard, and as such have stagnated a bit.
3.) The mobile "revolution" has shifted the emphasis of CPU design from desktop and high performance to mobile and low power.
These three combined have resulted in Intel performance gains since Nehalem to be very lackluster.
October 2016? What the heck does AMD plan to do for the next year?
One thing is to work with GF to get their 14nm process ready for production runs at the frequency that AMD thinks it needs to be competitive.
I'm switching to AMD for my main rig if it has Sandy Bridge IPC. I'm an AMD fanboy for life!!!!!
I went from Pentium to Duron to Athlon XP to Athlon 64 to Athlon 64x2 to i5 Nehalem.
Whoever offers the best "bang for the buck" always gets my money.
Maybe it'll be TSMC.