Is it time to jump?

I still like AMD and that is what I am using in my Desktop, Laptop and Tablet PC. (At least at this time.) What I figure is that AMD really started hurting when Hector Ruiz (Ruiniz?) sold off AMD's own manufacturing plants.

To me, that would be like FORD selling off all their plants and then getting GM to build their vehicles for them. Mr. Ruiz should have ended up in the same boat as his AMD run company did, not being able to float off in his golden parachute. :mad:
 
I don't believe steamroller is a completely new architecture. I realignment of resources, yes. Shifting of focus what what matters, again yes. But from my understanding it will still use modules and shared FPU. Kind of like the changes from the Radeon 4000 series to the 5000 series. Resources were shifted around and improved. Far more comprehensive than a tweak, but not a 100% redesign either.

I 'm expecting steamroller to have a 30% performance jump increase in IPC over Vishera with along with huge gains in efficiency.

AMD will not be the leader, but they should be back in the game.
 
I still like AMD and that is what I am using in my Desktop, Laptop and Tablet PC. (At least at this time.) What I figure is that AMD really started hurting when Hector Ruiz (Ruiniz?) sold off AMD's own manufacturing plants.

To me, that would be like FORD selling off all their plants and then getting GM to build their vehicles for them. Mr. Ruiz should have ended up in the same boat as his AMD run company did, not being able to float off in his golden parachute. :mad:

It wasn't a bad move in their case. It really wasn't. They couldn't handle developing manufacturing and fabrication along side design. The added control is a benefit no doubt, but it isn't what's important right now.
 
Wow, this turned in to quite the discussion, but Im glad that it did. There is a lot of good, healthy and somewhat heated discussion here.

One thing that I forgot to mention is that I dont only do gaming. I also do visual effects in After Effects/Vegas Pro, and some heavy photoshop.

The way I see it is this...

AMD FX-8150 = $199
Crosshair V Formula = $219
Total=$418

Intel i5 2500k = $169 (on sale at Microcenter)
Sabertooth Z77 = $239
Total = $408

For a scratch builder, the choice is clearly the 2500k, but lets look at the numbers with the 2700k i7

Intel i7 2700k = $289 (On sale at Microcenter)
Sabertooth Z77 = $239
Total = $528
Delta with AMD = +$110

so for $10 less you can get the 2500k which performs better all around, or for $110 more you can get the 2700k which is absolutely a no brainer when it comes to AMD vs Intel.

I think for me personally, the fact that I already have the Crosshair V Formula only helps the fact that I can sell it and more than offset the cost difference for the 2700k and sell the 965 for 40 or 50.

The bottom line is that in todays lineup, I have no idea why anyone would go AMD, and believe me when I tried to talk my 2 customers who went AMD this week out of it.. but then again, as a system builder, I have to offer choice. *shrug*
 
If AMD gets the deals to be making the graphics for the XBOX 720, new Nintendo and PS4, which is looking very likely, won't that give them back some edge?
 
If AMD gets the deals to be making the graphics for the XBOX 720, new Nintendo and PS4, which is looking very likely, won't that give them back some edge?

Only in profits. Not necessarily technology.
 
If AMD gets the deals to be making the graphics for the XBOX 720, new Nintendo and PS4, which is looking very likely, won't that give them back some edge?

In gaming performance over nVidia, sure. GPUs differ from CPUs in that programming for them is closer-to-metal than CPUs, so things like Gaming Evolved and TWIMTBP offer that closer-to-hardware relationship to developers and in turn it equates to better performance for a specific type of GPU architecture over another. In general, nVidia has been far more willing to do the extra leg work to make certain their GPU architecture gets the extra treatment than AMD has, but with consoles looking likely to go AMD this round it will translate into better performance for AMD architectures that look a lot like the console chips.

The profits for AMD from console sales aren't that great if you at it per-chip sold, so their income will depend heavily on quantity. Remember, consoles have to remain cheap to be viable gaming platforms.

edit - this depends heavily on the GPU architecture and for what platform the game is targeted. For instance, there can still be plenty of TWIMTBP games if they're PC only titles. The overwhelming majority of games would still favor AMD's GCN architecture, or whatever architecture they end up using for that specific console.
 
Last edited:
I hate to say it, but I use AMD only because I want variety with my computers. My #2 machine is an aging 1090t. If it dies, I'd move my 2600k and p67 over and get a 3770k and Z77 board for my main. I heard the IPC on Trinity is better, so this may be good for the upcoming PileDriver based FX chips....As long as they perform at least as well as the old Nehalems(this really isn't asking for much...Nehalem is ANCIENT), I'll see AMD as viable again.
 
GPUs differ from CPUs in that programming for them is closer-to-metal than CPUs,

Huh, Really?

And here I was thinking that games were written strictly to a DX or OpenGL abstraction layer isolating the hardware itself completely from the code...

This is interesting. Do you know where I can read up more on it?
 
why anyone would do anything but buy intel for their machines is beyond me
 
He balked on my kepler vs. GCN in gaming question, though. The other ones he actually answered. Mainly because he and I both know the answer to that :p If you focus more on compute you lose some of the pure gaming power when designing a GPU architecture. It's part of the reason why GCN is behind the GK104.

Here's the snippet

WinZip and AMD have been promoting their joint venture in implementing OpenCL hardware accelerated compression and decompression. Owning an AMD GPU I appreciate it. However, it's been reported that WinZip's OpenCL acceleration only works on AMD CPUs. What is the reasoning behind this? Isn't it hypocritical, given AMD's previous stance against proprietary APIs, namely CUDA, that AMD would then support development of a vendor specific OpenCL program?

The OpenCL version of Winzip has been optimized on AMD GPUs and achieves significant performance gains. While OpenCL is not vendor specific optimizations on any application are essentially vendor specific since they depend on the microarchitecture of eachvendor. We worked closely with Winzip to get these optimizations in. We have the most mature OpenCL implementation currently and even then we just managed to get the QA done before WinZip's launch date. I am sure that they will be coming out with OpenCL optimized versions on Intel and Nvidia soon -- you should ask them). That is in fact the beauty of OpenCL – one code base gives functional portability across vendor platforms and optimizations are the only components that need to be scheduled. So yes this is in line with our embrace of open and industry standards.

This may be related to the above situation. Even with standardized, cross-platform, cross-vendor APIs like OpenCL, to get the best performance, developers would need to do vendor specific, even device generation within a vendor specific optimizations. Is there anything that can be done whether at the API level, at the driver level or at the hardware level to achieve the write-once, run-well anywhere ideal?

Device specific optimizations will always have a beneficial impact on performance. This is true even with CPUs. While differences between GPUs are more dramatic, this is due to the fact that today's GPUs are designed to excel at graphics while compute is a secondary consideration. Reluctance to spend more chip area on compute results in having many device specific performance "cliffs". For example, VLIW instructions, 64 thread wavefronts, and the need for coalesced accesses to memory. As GPUs are increasingly used for compute, and as it becomes possible to add yet more transistors, these "cliffs" will continue to disappear. Advances in compilers will also help.

In essence, despite the uniformity of DX/openGL APIs, the architectural differences in GPUs tend to be quite vast and far more different than your regular IB v. Bulldozer architecture is. It's the reason we have GE vs. TWIMTBP initiatives in the first place. If AMD delivers a roughly similar architecture in their discrete GPUs/APUs on laptops/desktops as their consoles then they'll hold this advantage. Frankly, it's worth more than the money they'll make off of each console. Even if they break even it's still a massive notch for them against nVidia.
 
So iyo is that why the 670/680's are so weak when it comes to OpenCL or the compute performance thing?

(Because Nvidia doesn't want the masses to understand its benefits, so therefore the masses see Open CL as worthless and it can hurt their competitor if the status quo don't know about its benefits?)

On topic IMHO AMD CPU OC'd+ single strong GPU OC'd+ fast SSD SATA III+ Fast Nic is still an awesome setup ;) For SLI/CFX setups, yeah better to go Intel it sounds like for a CPU. ( I still have to get an SSD and current gen GPU before I jump ship to Intel. That and an extra 500 buckaroos might do the trick :D In the meantime I'll still be getting very high K/D ratios with my AMD setup and no hax, I swear :)
 
Wow, this turned in to quite the discussion, but Im glad that it did. There is a lot of good, healthy and somewhat heated discussion here.

One thing that I forgot to mention is that I dont only do gaming. I also do visual effects in After Effects/Vegas Pro, and some heavy photoshop.

The way I see it is this...

AMD FX-8150 = $199
Crosshair V Formula = $219
Total=$418

Intel i5 2500k = $169 (on sale at Microcenter)
Sabertooth Z77 = $239
Total = $408

For a scratch builder, the choice is clearly the 2500k, but lets look at the numbers with the 2700k i7

Intel i7 2700k = $289 (On sale at Microcenter)
Sabertooth Z77 = $239
Total = $528
Delta with AMD = +$110

so for $10 less you can get the 2500k which performs better all around, or for $110 more you can get the 2700k which is absolutely a no brainer when it comes to AMD vs Intel.

I think for me personally, the fact that I already have the Crosshair V Formula only helps the fact that I can sell it and more than offset the cost difference for the 2700k and sell the 965 for 40 or 50.

The bottom line is that in todays lineup, I have no idea why anyone would go AMD, and believe me when I tried to talk my 2 customers who went AMD this week out of it.. but then again, as a system builder, I have to offer choice. *shrug*

Not a completely fair comparison. The FX-8120 is $169 and is the same thing as a 8150 with just a couple clicks in the BIOS. And the 990FX Sabertooth is $189 so $60+ cheaper than the Intel Sabertooth. And you can't count the Microcenter price on the 2500K cause very few people live near one, plus you'll have sales tax. So you've got a $170 FX-8120 compared to a $220 i5 and a $190 AM3+ Sabertooth compared to a $250 Intel Sabertooth. So that's a $110 difference. That's a lot better deal when you compare prices equally.
 
Last edited:
So iyo is that why the 670/680's are so weak when it comes to OpenCL or the compute performance thing?

(Because Nvidia doesn't want the masses to understand its benefits, so therefore the masses see Open CL as worthless and it can hurt their competitor if the status quo don't know about its benefits?)

On topic IMHO AMD CPU OC'd+ single strong GPU OC'd+ fast SSD SATA III+ Fast Nic is still an awesome setup ;) For SLI/CFX setups, yeah better to go Intel it sounds like for a CPU. ( I still have to get an SSD and current gen GPU before I jump ship to Intel. That and an extra 500 buckaroos might do the trick :D In the meantime I'll still be getting very high K/D ratios with my AMD setup and no hax, I swear :)

I'm not sure I totally understand you, but here goes.

Intel has the faster CPUs, the GPU can be anything, Intel's drive controller is faster than AMD's and more flexible, with the only downside being that it supports fewer SATA III 6Gb/s devices, and boards with Intel NICs are absolutely better performers than those equipped with anything else. Broadcom being the only one that's even close.

Single GPU or multiGPU, Intel produces the faster and more advanced overall platform. Period, end of story. The only way it ever makes sense to buy AMD is if performance is secondary to cost. In which case, depending on what features you really need and whether or not you have a Microcenter near you, the price vs. performance ratio isn't always so clear cut. People act like every Intel LGA1155 board is expensive and that their CPUs are vastly more costly. This is not always the case. Feature for like features AMD boards are cheaper, but if you won't leverage half those features, who cares? Buy the better performing platform.
 
Last edited:
So iyo is that why the 670/680's are so weak when it comes to OpenCL or the compute performance thing?

It's mostly because nVidia half-asses their openCL/openGL performance because it takes away from CUDA. CUDA performance also sucks though, but that's because if it didn't then it would take away their sales from Tesla GPUs which have much higher profit margins. It's shitty, yea, but nVidia purposely handicaps their GPUs so they can make more money from HPC.

This isn't necessarily bad, though. For mainstream GPGPU use it is, but for HPC this makes sense because nVidia can then offer two distinct products based on the same architecture. So they can concentrate on higher clocks and lower bus width and less ram for gaming (GTX 680) and still beat AMD but also kick AMD's ass in compute with 512bit bus widths and 6GB ram and CUDA support. AMD tried to appease both markets at once but it's cost them a bit of performance for gaming and they can't quite catch up in compute, but they're definitely getting close and it's also much cheaper. openCL/openGL performance AMD wins but mostly because nVidia can't risk losing CUDA. Technically, CUDA will always be faster because it's built around the architecture itself whereas openCL is cross-platform (it's really Apple's baby). So with CUDA you can squeeze out that much more performance and better program/optimize your software such that it fits your hardware perfectly. AMD doesn't yet have a proprietary coding tool like that and they're relying on openCL being popularized in mainstream first and seeping into HPC. What AMD has going for it atm is savings in price but thus far HPC users have steered away from openCL and selected to opt with the more expensive Tesla/CUDA platforms.

There's a good link that explains most developers' feelings toward this issue in HPC but I can't find the link. When I do I'll post it here.
 
Last edited:
I'm not sure I totally understand you, but here goes.

Intel has the faster CPUs, the GPU can be anything, Intel's drive controller is faster than AMD's and more flexible, with the only downside being that it supports fewer SATA III 6Gb/s devices, and boards with Intel NICs are absolutely better performers than those equipped with anything else. Broadcom being the only one that's even close.

Single GPU or multiGPU, Intel produces the faster and more advanced overall platform. Period, end of story. The only way it ever makes sense to buy AMD is if performance is secondary to cost. In which case, depending on what features you really need and whether or not you have a Microcenter near you, the price vs. performance ratio isn't always so clear cut. People act like every Intel LGA1155 board is expensive and that their CPUs are vastly more costly. This is not always the case. Feature for like features AMD boards are cheaper, but if you won't leverage half those features, who cares? Buy the cheaper board.

AMD is also an option for those who are currently on an AMD platform like me. Yeah I'd love to switch over to Intel but it would cost me $400+ to do so. Since all I do is game at 1920x1080 with a single GPU, it makes absolutely no sense unless I've just got the cash to spare.

Also, AMD is cheaper. Like I posted above. $170 for a 8120 vs $220 for a 2500K vs $250 for a 3570K. Not to mention the $60 premium I'd have to pay for an Intel Sabertooth (once you have a Sabertooth, you just can't go back ;) ).

And unless you've got $2000 invested in GPU's and monitors, you're simply not going to see a real world difference in 99% of games with an AMD vs Intel system at 2560x1600 or below with a sub $500 video card.

Anand had a line in a recent review: "there are no such things as bad products, just bad prices." I think that applies here. Intel is faster but that doesn't mean AMD sucks especially when they're a good bit cheaper. There is a place for them and you can have a great performing rig with a FX in the socket.
 
It's mostly because nVidia half-asses their openCL/openGL performance because it takes away from CUDA. CUDA performance also sucks though, but that's because if it didn't then it would take away their sales from Tesla GPUs which have much higher profit margins. It's shitty, yea, but nVidia purposely handicaps their GPUs so they can make more money from HPC.

This isn't necessarily bad, though. For mainstream GPGPU use it is, but for HPC this makes sense because nVidia can then offer two distinct products based on the same architecture. So they can concentrate on higher clocks and lower bus width and less ram for gaming (GTX 680) and still beat AMD but also kick AMD's ass in compute with 512bit bus widths and 6GB ram and CUDA support. AMD tried to appease both markets at once but it's cost them a bit of performance for gaming and they can't quite catch up in compute, but they're definitely getting close and it's also much cheaper. openCL/openGL performance AMD wins but mostly because nVidia can't risk losing CUDA. Technically, CUDA will always be faster because it's built around the architecture itself whereas openCL is cross-platform (it's really Apple's baby). So with CUDA you can squeeze out that much more performance and better program/optimize your software such that it fits your hardware perfectly. AMD doesn't yet have a proprietary coding tool like that and they're relying on openCL being popularized in mainstream first and seeping into HPC. What AMD has going for it atm is savings in price but thus far HPC users have steered away from openCL and selected to opt with the more expensive Tesla/CUDA platforms.

There's a good link that explains most developers' feelings toward this issue in HPC but I can't find the link. When I do I'll post it here.

Oh yeah I forgot about Tesla, now that makes sense. Good call.
 
Lets quantify things. I AM an AMD fanboi but they have made to many incorrect decisions.

My 8150 4.5GHz bulldozer with two 2GB 6950s did 55fps at Medium in BF3.
My 3570k (stock clock for now) with two 670s does 65fps at High in BF3 and only draws 440w from the wall. ITS SICK.

I love AMD and when I build machines for others or recommend laptops, if not for gaming, I recommend AMD all day long. As soon as high end gaming comes into play I reluntantly tell them to go IB these days.

Bulldozer is a fail and pilediver is just a little less stinky poo at this point. If steamroller comes out and rocks the house (?) I will gladly switch back.
 
AMD is also an option for those who are currently on an AMD platform like me. Yeah I'd love to switch over to Intel but it would cost me $400+ to do so. Since all I do is game at 1920x1080 with a single GPU, it makes absolutely no sense unless I've just got the cash to spare.

Also, AMD is cheaper. Like I posted above. $170 for a 8120 vs $220 for a 2500K vs $250 for a 3570K. Not to mention the $60 premium I'd have to pay for an Intel Sabertooth (once you have a Sabertooth, you just can't go back ;) ).

And unless you've got $2000 invested in GPU's and monitors, you're simply not going to see a real world difference in 99% of games with an AMD vs Intel system at 2560x1600 or below with a sub $500 video card.

Anand had a line in a recent review: "there are no such things as bad products, just bad prices." I think that applies here. Intel is faster but that doesn't mean AMD sucks especially when they're a good bit cheaper. There is a place for them and you can have a great performing rig with a FX in the socket.

Just because you won't do less than a Sabertooth doesn't mean that there aren't viable cheaper options. If you only need the most basic features, there is no sense in buying a board with tons of features you will never use just because it's the same price on the AMD side. If a simple board gives you what you need but gives you the option for Intel processors, I'd take it. Additionally, it's hard to quantify what the value of more performance is. For me $110 difference or even more for a much better performing product is absolutely worth while and justified.
 
Last edited:
Not a completely fair comparison. The FX-8120 is $169 and is the same thing as a 8150 with just a couple clicks in the BIOS. And the 990FX Sabertooth is $189 so $60+ cheaper than the Intel Sabertooth. And you can't count the Microcenter price on the 2500K cause very few people live near one, plus you'll have sales tax. So you've got a $170 FX-8120 compared to a $220 i5 and a $190 AM3+ Sabertooth compared to a $250 Intel Sabertooth. So that's a $110 difference. That's a lot better deal when you compare prices equally.

Heck, if you factor in Microcenter pricing, you can get an FX-8120 AND a Gigabyte GA-990FXA-UD3 for $244.98 as a bundle. That's quite a deal.
 
Here's the link I was talking about

http://vr-zone.com/articles/from-gt...-kepler-in-real-gpgpu-deployments-/15903.html

To us, all of this sounds ideal for AMD, and our sources went on to say that the real limitation is in the way how AMD positions its products. “AMD went for cheap and affordable”, said our source. "They don’t have anything for the servers and that’s what guts them. Our clients go AMD 7970, but then we have to go with a 5U chassis and put up to eight Radeon HD 7970 cards, but we cannot compete on space optimized 1U and 2U, since AMD has no purpose-built (GPGPU) products for that."

And here lies the sad truth. GPGPU is turning to be a big business for NVIDIA, earning more money than AMD is able to pull from their consumer endeavours. BUT, since there are no contemporary server-grade parts from AMD and “if the company requires a 2U or 1U, AMD solution is SOL and we go NVIDIA”.

The quotes above come from several C-level executives, and you can see how similar they are.

Here’s an advice for AMD - if you guys want to make money using your existing technology, how about qualifying a HD 7970 with 6GB of ECC memory and call it FireStream 7970? Just that there aren’t any surprises when next round of financial results come in and NVIDIA once more demolishes AMD by effectively making a technology which is inferior in simple GPGPU apps.
 
This is old news cause I remember hearing this same thing several months ago but for some reason it showed up on Tom's Hardware today at 5PM EST. Anyways, thought it was relevant.

AMD backing out of Speed Wars with Intel

There is no interest at AMD to continue a processor war with Intel that has lasted decades, but only rewarded AMD with occasional superiority. “That era is done,” Rory Read said in an interview and added, “There’s enough processing power on every laptop on the planet today.”
Again, I know this was first out months ago but if Read's still running around saying this, makes me kinda sad. But it could still be that while AMD is giving up trying to be faster than Intel, maybe they can at least be competitive. Trinity is a good increase over Llano and Vishera looks like it might be a decent upgrade as well. If Steamroller makes a big jump and is priced right, AMD can still be competitive while not trying to be faster than Intel.
 
Last edited:
with the amount of resources amd has to develop chips one can only hope that that staying competitive is what they aim to do because if not they will become irrelevant and we'll only have one real choice for computing
 
Honestly, who cares about that score. Play some games, run some applications, don't just use an arbitrary score like 3DMark. Get a feel for which is snappier, which is smoother, which one you actually like using.

The one that feels smoother and snappier is the i5/i7's.
 
I jumped ship from AMD, trust me, you won't regret it. Intel is so far ahead now it isn't even funny, its sad.
 
Pelo: I think that piece is wrong where they state it is big business it is a business but big I doubt it. It is a niche segment of the market at best. Something like making GPU for a console is rather large business.

If Nvidia could make such a sick amount of money by making only GPGPU devices why are they still not doing that ?
 
They are... Teslas? Did you bother to read the article?

In fact HPC is expanding at a larger rate than discrete video cards for gaming. It's why AMD went with the GCN architecture in the first place; part compute/part gaming in an attempt to appeal to both markets. The difference is they haven't offered HPC-grade hardware yet.
 
I jumped ship from AMD, trust me, you won't regret it. Intel is so far ahead now it isn't even funny, its sad.

Agreed.

It is sad but true.

I am an AMD fan, but even I can't justify staying on an AMD platform anymore., at least not for my main "enthusiast" rig. I still use AMD CPU's for family non-gaming rigs, HTPC's and servers, and if I had my own laptop that wasn't issued by work it would likely be AMD Fusion based.

I just can't justify having an AMD CPU in my main rig anymore, not considering how far behind they are.

On purely single threaded workloads at the same clock speed Ivy Bridge now outperforms BD between 75% and and 87% depending on the benchmark. It is very sad.

If AMD could increase their clock speed and compete that way, it would be one thing, but they apparently can't. The top turbo speed of an FX-8150 is only 300mhz faster than the top turbo speed of a 3770K (4.2 vs. 3.9)

If we factor in overclocking, in order to match a 3770K at 4.7Ghz an FX-8150 would need to hit between 8.2 to 8.7Ghz. That simply isn't going to happen.

Now, being an AMD fan as I am, they wouldn't have to beat or tie Intel for me to go AMD. They would however have to come close enough that I am not giving up anything significant by going AMD. If AMD came within 10-15% of Intel, I'd go AMD all the way, but at these massive losses, I just can't justify it anymore, especially since my favorite game (and only one I play) right now is particularly CPU intensive, and is CPU limited to the point where I can't find an AMD CPU that will keep it at 60+ fps at all times regardless of how high it is overclocked.

I already invested a bunch of money in my current CPU and motherboard, and likely will use them for a while, but if AMD can catch up to where their single threaded performance is within 10-15% of Intel ( (Intel perf - AMD Perf) / AMD Perf ) for the next generation, I'm all AMD.
 
Last edited:
The biggest dilema for me now is i5 2500k or i7 2600k. Reading the quotes by AMDs own representitives makes me now 100% positive that they are not interested in improving my experiences in what matters most to me. Therefor, they made the decision for me.
 
Zarathustra[H];1038759412 said:
Heck, if you factor in Microcenter pricing, you can get an FX-8120 AND a Gigabyte GA-990FXA-UD3 for $244.98 as a bundle. That's quite a deal.

For what its worth, it would likely never be a desktop machine for me (8 cores is excessive, and per core performance is too low), but I have been considering this combo deal to upgrade my aging server which runs ESXi and a few VM's. The 8 cores will really come in handy here, and per core performance won't be very important for what I am doing.
 
Yes sadly Intel is the way to go now, but I question because of AMD no longer participating in a speed war that Intel will no longer have the itch to "really" speed up any future processors. The enthusiast market will become stagnate IMHO.:confused:
 
Yes sadly Intel is the way to go now, but I question because of AMD no longer participating in a speed war that Intel will no longer have the itch to "really" speed up any future processors. The enthusiast market will become stagnate IMHO.:confused:

We will have to wait and see.

It may already be happening. The Ivy Bridge CPU's really aren't much faster than Sandy Bridge, and why should Intel bother when AMD still has nothing that can challenge Sandy?
 
Zarathustra[H];1038763178 said:
We will have to wait and see.

It may already be happening. The Ivy Bridge CPU's really aren't much faster than Sandy Bridge, and why should Intel bother when AMD still has nothing that can challenge Sandy?

Yep exactly.
 
Zarathustra[H];1038763178 said:
We will have to wait and see.

It may already be happening. The Ivy Bridge CPU's really aren't much faster than Sandy Bridge, and why should Intel bother when AMD still has nothing that can challenge Sandy?

The refresh cycles are not usually that big a jump anyway. I don't see that as a sign of anything. Intel hasn't really had to seriously compete with AMD since the introduction of the Core 2 Duo.
 
I would say if you're honest with yourself AMD can be a viable solution. For me it makes sense primarily because I don't game as much as I used to. However, I do a lot of video encoding, graphic work, and VM test beds and on that front AMD is competitive with Sandy.

When BD came out I was particularly interested in it's x264 performance. Considering it was competitive I looked at leveraging AMD's lower prices in it's server line. So I went from a i7 920 to a dual socket C32 system a couple of weeks ago. The end results were staggering. The system cost me less than what a 3960x system would cost me but it encodes video almost as well.
 
I would say if you're honest with yourself AMD can be a viable solution. For me it makes sense primarily because I don't game as much as I used to. However, I do a lot of video encoding, graphic work, and VM test beds and on that front AMD is competitive with Sandy.

When BD came out I was particularly interested in it's x264 performance. Considering it was competitive I looked at leveraging AMD's lower prices in it's server line. So I went from a i7 920 to a dual socket C32 system a couple of weeks ago. The end results were staggering. The system cost me less than what a 3960x system would cost me but it encodes video almost as well.

That's pathetic.
 
That's pathetic.
Wow Dan your faith in me sucks...lol.

I knew someone would ask about the almost, but I decided to say almost for brevity. However, I hoped the response would have been more inquisitive. It has more to do with the limitations of the x264 encoder than anything else. For every additional thread added to an encode the effect on performance is halved.

During an encode my system utilization is at 50%. In order to max the system out I encode 2 files at the same time which has a negligible impact (meaning minutes not hours) on the first encode. What does this mean? If you need numbers my old system could only do 1 encode at a time at 100% utilization. DVD's encoded at somewhere between 1 - 2 hours (high profile) and BR titles were anywhere from 3 - 5 hours. This is now 30- to 45 minutes for 2 DVDs or 2 BR titles at no more than 3 hours. I'm sorry but I'm loving that performance every day even on Sundays. I ran the numbers before I committed. I wouldn't spend that kind of money if it wasn't going to perform. I knew what my workload was before I bought it. You have to consider the resources available for highly threaded workloads.

In addition that CPU costs $1000+ vs the $700 I spent I think i can live with the 30 fps deficit and 300 more dollars in my pocket. Just to be thorough, there's a 30 fps deficit in my favor between a Extreme 980x at stock and this system.
 
Last edited:
I would say if you're honest with yourself AMD can be a viable solution. For me it makes sense primarily because I don't game as much as I used to. However, I do a lot of video encoding, graphic work, and VM test beds and on that front AMD is competitive with Sandy.

When BD came out I was particularly interested in it's x264 performance. Considering it was competitive I looked at leveraging AMD's lower prices in it's server line. So I went from a i7 920 to a dual socket C32 system a couple of weeks ago. The end results were staggering. The system cost me less than what a 3960x system would cost me but it encodes video almost as well.

Yeah, but a 3960x is not a very good comparison. At $1000 we know it is an overpriced CPU, so doing a cost/performance comparison to it is not really that useful.

Compare it to a 3930k instead. Even better, an overclocked 3930k. In Cinebench 11.5 I score 13.4 with mine at 4.6ghz. I don't have much to compare this to, but it's 70% faster than a 12 core Opteron 2435... I'd compare encode speeds, but I never really encode anything, and don't even know how to test the speed repeatedly.

Rendering and encoding is still an area where AMD can mostly keep up with Intel due to the higher core counts for the price. However, if you consider the fact that PC's are multi-purpose machines, why would you choose a system that can just barely tie its competitor in one very specific task, yet lags behind it SIGNIFICANTLY in almost everything else?

Wouldn't one be better served by the competitor that can do everything well?

I'll concede that for VM's AMD's FX line is great. This is one area where extra cores really are useful, more so than individual core speed in most cases. I am planning on picking up an FX-8120 for my ESXi server, as 8 cores for a CPU and motherboard combo for just over $200 is hard to beat. This at best represents a a VERY niche user base though. People who have ESXi or other VM servers, or ever even use a workstation-type VM are probably in the less than 1/1000 of one percent of computer users category, so it's a tough advantage to trumpet all over the place.
 
Back
Top