AMD faces class action suit over Bulldozer missrepresentation

[SurvivorNVL]

Haha. I'm stuck between systems engineering and comsci but I'm gravitating to the one that is stupid hard. Computer systems-.-

Good luck on that.  When in doubt--I'll probably get a job at a bank and work my way up to owning every Suntrust in the country.  It's not impossible.  Just have to use our imaginaaaation.

[TechGod]

Good luck on that. When in doubt--I'll probably get a job at a bank and work my way up to owning every Suntrust in the country. It's not impossible. Just have to use our imaginaaaation.

the best plan.

[LAwLz]

H.264 isn't lossless for the record without use of the maximum bit rates. Mine's so close you can't even tell unless you look at the bytes themselves or line up 2 of the same frame and check pixels with a magnifying glass.

Ehm, yes? That's exactly what I said here:

Oh so it was not really a lossly compression (except some errors)? I am far less impressed now.

Lossy compression is far harder to write, especially if you want to get anywhere near the size:quality of for example H.264 (uses a lot of complex predictions, filters and other techniques). You can find more info about it on this website (login: fj@mailinator.com Password: Qwerty123).

 

 

 

And why are you so much less impressed? I'm within 2% of your ratio with hardly any effort. Now, is it streamable? Not unless someone writes a container for it. Can you skip to wherever you want in the movie? Not easy since I don't use I-Frames. That said, VP9 and H.264 are not implemented well from a data and instruction-level parallelismstandpoint, or the compilation is garbage. If you're telling me it takes thousands of operations per bit to decompress a video or encode it with high retention of quality and a high compression ratio, I'm going to laugh at your expense. It's bullshit.

It is far less impressive because with lossless there is no flexibility. With for example H.264 they have to think about how to lose as little quality as possible while still keeping the file as small as possible. You only have to do the latter with lossless. One major part of H.264 is the deblocking filter. You could not even make a thing like that for your format because then it would not be lossless anymore. I doubt you implemented had anything as advanced as that or motion prediction

 

Laugh all you want, but the sad truth is that you are only laughing because you are ignorant about how the codecs work and why they require so much processing power. Try it yourself. Get HandBrake and encode something with the placebo preset. Maybe throw in some filters as well if you want. Or just read the documentation I linked earlier and you might realize how very complex some of these codecs are and why they require so much processing power. Each block in each frame gets processed several times in order to get the maximum amount of compression with the least amount of quality loss.

I mean, I can link you to what kind of performance I get if you want (already did it earlier in the thread, 3 FPS). If you are going to blame the encoder for the low performance then I will call you an idiot because you don't even understand the basics of the codec and could probably not even write an encoder that's 1/4 as good as it.

 

 

 

Because unless the CPU is busy, you're wasting energy putting it on the GPU.

Ehm, no? At least not with HSA. In fact, you are wasting energy doing it in the CPU if the GPU can do it faster and more efficiently.

[patrickjp93]

Ehm, yes? That's exactly what I said here:

 

 

 

It is far less impressive because with lossless there is no flexibility. With for example H.264 they have to think about how to lose as little quality as possible while still keeping the file as small as possible. You only have to do the latter with lossless. One major part of H.264 is the deblocking filter. You could not even make a thing like that for your format because then it would not be lossless anymore. I doubt you implemented had anything as advanced as that or motion prediction

 

Laugh all you want, but the sad truth is that you are only laughing because you are ignorant about how the codecs work and why they require so much processing power. Try it yourself. Get HandBrake and encode something with the placebo preset. Maybe throw in some filters as well if you want. Or just read the documentation I linked earlier and you might realize how very complex some of these codecs are and why they require so much processing power. Each block in each frame gets processed several times in order to get the maximum amount of compression with the least amount of quality loss.

I mean, I can link you to what kind of performance I get if you want (already did it earlier in the thread, 3 FPS). If you are going to blame the encoder for the low performance then I will call you an idiot because you don't even understand the basics of the codec and could probably not even write an encoder that's 1/4 as good as it.

 

 

 

Ehm, no? At least not with HSA. In fact, you are wasting energy doing it in the CPU if the GPU can do it faster and more efficiently.

You said lossly. I couldn't tell which you meant.

 

There's plenty of ways to get lossless and still stay small. I could add I-frames to be able to jump around and it would improve the quality more to have more pre-defined full frames. I could also adjust the granularity back and lose a bit of the compression ratio. I created something pretty good for a total amateur.

 

I'm saying the computational power they use is wasted because it can be done better, which is provable mathematically even without an implemented program. All graphics other than ray tracing are embarrassingly parallel. Video is reducible to graphics and is reducible to embarrassingly parallel compute. Therefore, if you're telling me it takes thousands of manipulations per bit, I'm saying bullshit. It's not necessary.

 

The GPU can't do it more efficiently if there's any branching or sequential processing involved (which H.264 has both of). You'd be better off with a Xeon Phi.

[LAwLz]

There's plenty of ways to get lossless and still stay small. I could add I-frames to be able to jump around and it would improve the quality more to have more pre-defined full frames. I could also adjust the granularity back and lose a bit of the compression ratio. I created something pretty good for a total amateur.

Depends on what you classify as "small". FLAC can create some pretty small files compared to PCM (uncompressed audio), but it will still be huge compared to for example Opus or MP3 at reasonably low bitrates. Same with video. Lossless compression might be small compared to the RAW video, but you will very rarely be able to get anywhere near the small file size you can get with lossy compression.

I haven't seen your codec so I can't tell if it's "pretty good for a total amateur", but you brought it up as some huge merit when it is most likely not anywhere near as good as for example x264. It sounds rather pathetic to me. Kind of like if I were to say Battlefield 4 is poorly programmed and then bring up my battleship game I wrote in high school as a big merit to prove that I know what I am talking about. "My battleship game ran much better than Battlefield 4 and I did not have as many issues with my netcode. Battlefield 4 is just poorly programmed and I will laugh at it because I am so awesome". That's what you sound like right now.

 

 

I'm saying the computational power they use is wasted because it can be done better, which is provable mathematically even without an implemented program. All graphics other than ray tracing are embarrassingly parallel. Video is reducible to graphics and is reducible to embarrassingly parallel compute. Therefore, if you're telling me it takes thousands of manipulations per bit, I'm saying bullshit. It's not necessary.

Did I just stroke out? Did you just stroke out? Someone must have had a stroke because that sentence is just gibberish.

I am saying that encoding video requires a lot of processing power. That's what I am saying, and it does. There is a huge amount of variables that needs to be calculated, tested, recalculated and compared for every single frame.

It is a problem that lends itself very well to parallelization, which is why programs such as x264 runs beautifully even if you throw like 50 cores at it. That does not reduce the amount of work that has to be done though. You still have to go through every macroblock multiple times and try and figure out the most optimal way to compress it without losing that much visual quality. I strongly recommend you actually read the H264 documentation and explanation I linked earlier. Right now it seems like you are still stuck on the idea that modern video codecs work on a pixel-by-pixel basis, which they don't. If they did they would require a lot less processing power, but their compression ratio would also be absolutely terrible compared to what we get with things such as motion vectors.

 

 

The GPU can't do it more efficiently if there's any branching or sequential processing involved (which H.264 has both of). You'd be better off with a Xeon Phi.

Yeah, because I am going to get a Xeon Phi for my computer. How much does that cost again? Like 2000 dollars? When my friend wants to livestream something I will tell him not to use the GPU to encode the video, I will tell him to get a 2000 dollar card and use that instead. I will admit that the result will most likely be a lot better (because the quality you get from ATI Stream and CUDA encoders is terrible compared to doing it on the CPU) but I don't think he will be willing to spend that kind of money.

[patrickjp93]

Depends on what you classify as "small". FLAC can create some pretty small files compared to PCM (uncompressed audio), but it will still be huge compared to for example Opus or MP3 at reasonably low bitrates. Same with video. Lossless compression might be small compared to the RAW video, but you will very rarely be able to get anywhere near the small file size you can get with lossy compression.

I haven't seen your codec so I can't tell if it's "pretty good for a total amateur", but you brought it up as some huge merit when it is most likely not anywhere near as good as for example x264. It sounds rather pathetic to me. Kind of like if I were to say Battlefield 4 is poorly programmed and then bring up my battleship game I wrote in high school as a big merit to prove that I know what I am talking about. "My battleship game ran much better than Battlefield 4 and I did not have as many issues with my netcode. Battlefield 4 is just poorly programmed and I will laugh at it because I am so awesome". That's what you sound like right now.

 

 

Did I just stroke out? Did you just stroke out? Someone must have had a stroke because that sentence is just gibberish.

I am saying that encoding video requires a lot of processing power. That's what I am saying, and it does. There is a huge amount of variables that needs to be calculated, tested, recalculated and compared for every single frame.

It is a problem that lends itself very well to parallelization, which is why programs such as x264 runs beautifully even if you throw like 50 cores at it. That does not reduce the amount of work that has to be done though. You still have to go through every macroblock multiple times and try and figure out the most optimal way to compress it without losing that much visual quality. I strongly recommend you actually read the H264 documentation and explanation I linked earlier. Right now it seems like you are still stuck on the idea that modern video codecs work on a pixel-by-pixel basis, which they don't. If they did they would require a lot less processing power, but their compression ratio would also be absolutely terrible compared to what we get with things such as motion vectors.

 

 

Yeah, because I am going to get a Xeon Phi for my computer. How much does that cost again? Like 2000 dollars? When my friend wants to livestream something I will tell him not to use the GPU to encode the video, I will tell him to get a 2000 dollar card and use that instead. I will admit that the result will most likely be a lot better (because the quality you get from ATI Stream and CUDA encoders is terrible compared to doing it on the CPU) but I don't think he will be willing to spend that kind of money.

You used a straw man argument. Battleship and Battlefield 4 are not the same game, not even the same genre. What I'm comparing is purely my compression ratio (which is competitive even while near lossless), my total compression time, and my total decompression time. I already said it's not a fully usable codec since it lacks I-Frames, but I'm only 10% larger than the same file encoded to H.264. 

 

It's not gibberish. It's long-winded. I'm saying the power they use is wasted. It's lots of small, singular 32-bit manipulations instead of vectorized 64, 128, or 256-bit manipulations. You're leaving performance on the table while using up your "power"

 

No it doesn't. All it takes to get decent is the first frame and the deltas between each sequential frame after. Even in the worst case of lossless compression where you had to calculate every pixel of every frame, so 4 bytes per pixel, that's AVX 128 once per pixel. Exactly once: vector_end_pixel = vector_previous_frame_pixel + vector_delta. 2 million times per frame. R3 = R1 + R2. Send R3 to the color frame, R1 = R3, R2 = popq (pop quad), rinse and repeat. At 60fps thats 120 million, or 480MegaFlops. That's a joke even for a 1 GHz core of Nehalem. The designers of these need to get back to basics and start with vectorization instructions in mind. I don't care how much extra junk you added to compress it further than DCC, if you made the computations so complex a 100GFlop processor can't keep up, you dun goofed, especially since videos that small can benefit much more from having tighter cache coherency. Complexity =/= good design or implementation. It's solely complexity. I can make a facial recognition program that's stupidly complicated and performs like piss, or I can make a simple one that has competitive per-core performance with the best the U.S. government's got. It only takes about 500 lines of code to build the algorithm itself. The database interface and deletion of old negative matches will be more obviously.

 

There's no need to recheck variables if you can validate the rules in a general way. That's mistake #1. Cheapest Xeon Phi is around $400.

[Khvarrioiren]

 Nehalem.

 

Stop triggering me.

[LAwLz]

You used a straw man argument. Battleship and Battlefield 4 are not the same game, not even the same genre. What I'm comparing is purely my compression ratio (which is competitive even while near lossless), my total compression time, and my total decompression time. I already said it's not a fully usable codec since it lacks I-Frames, but I'm only 10% larger than the same file encoded to H.264.

[Citation Needed] on the same file only being 10% smaller with H.264 (you said you didn't test H.264, remember?).

And that is a really pointless test because H.264 was not designed for lossless compression. You might as well say JPEG is awesome because when you converted a photo to PNG it became huge. Different formats got different strengths and weaknesses. Barely anyone uses H.264 for lossless compression because it's stupid.

Your lossless codec is about as far away from H.264 as battleship vs Battlefield 4. Completely different benefits/drawbacks and completely different in how they work.

 

 

It's not gibberish. It's long-winded. I'm saying the power they use is wasted. It's lots of small, singular 32-bit manipulations instead of vectorized 64, 128, or 256-bit manipulations. You're leaving performance on the table while using up your "power"

What you are saying is that encoding H.264 should require barely any processing power at all. A few pages ago you said even an Atom processor should have more than enough power to do it in real time (just after you got codec and transcode mixed up):

Transcoding is an umbrella term under whic fall encoding and decoding.

I'm telling you it's a bad algorithm, because 1 minute of raw 4K footage is 3*3840*2160*60 =1.492 billion bytes with no initial compression at all(which would be obscene), yet it's still well within budget of a midrange Atom to compress in a single second. Delta Color compression both intra and inter-frame is possible with parallel coordinates which don't require you decompress one block of a frame before another. The next frame's block coordinates are relative to the first, but you can check all the transforms in parallel. We covered this in my algorithms class (which I aced). If codecs are having problems on chips with more than 10x the manipulations capacity relative to the data input size, it's a bad algorithm or a bad compiler setting where vectorized optimization a aren't being used, because a video is nothing more than a series of color transformations in pixel coordinates. It is the very definition of embarrassingly parallel per frame, and the move between frames can have the relative coordinate transforms done in embarrassingly parallel fashion as well.

 

What you fail to realize is that H.264 is a far more complex codec than you seem to think. I don't know how many times I have asked you to check the sources I have linked but you really should. Right now you are an idiot talking about things you really do not even understand the basics of, calling other peoples' code bad.

 

 

 

There's no need to recheck variables if you can validate the rules in a general way. That's mistake #1. Cheapest Xeon Phi is around $400.

Maybe recheck is the wrong wording. I was referring to the motion estimation. Each macroblock gets checked against several other blocks to try and find good matches. H.264 is actually so cleaver with motion estimation that it can apply motion vectors to a macroblock based on a FUTURE frame, not just a past one.

 

Anyway, maybe you should message Dark Shikari and tell him what terrible code he has written because according to you something that requires ~50 overclocked Sandy Bridge CPU cores with his program could be done with an Atom processor.

Oh and message Adobe, Ateme, Intel and all other companies that makes H.264 encoders as well. I am sure they will appreciate the laugh. And yes, Intel makes an h.264 encoder (included in IPP) and no, it's not even as good as x264 in terms of speed (or image quality).

[Guest]

So are we saying that they don't actually have 8 cores?? I was planning on buying an FX 8320, but after reading this article, I am being turned away from AMD ever so slowly. So to start a discussion, we are saying that they only have 4 cores? What about the FX 6000 series?? Do they only have 3 cores?? Someone help me pls. I plan on buying a new CPU and I am now not sure what to buy. I want something that is good for gaming, recording/streaming. As aforementioned, I was planning on buying the 8320, but have now been driven away from them. Another Question: why is it that in most games, the FX 8350 seems to compete so well with the 3rd and 4th generation CPUs? I notice that between the 8350 and the 4770K, in most games, there is only a 5-10 FPS difference. So are we saying that an AMD 4 core FX CPU is as powerful as an 8 core hyperthreaded i7?? Please help me...I really don't understand.

[JebKerman]

So are we saying that they don't actually have 8 cores?? I was planning on buying an FX 8320, but after reading this article, I am being turned away from AMD ever so slowly. So to start a discussion, we are saying that they only have 4 cores? What about the FX 6000 series?? Do they only have 3 cores?? Someone help me pls. I plan on buying a new CPU and I am now not sure what to buy. I want something that is good for gaming, recording/streaming. As aforementioned, I was planning on buying the 8320, but have now been driven away from them. Another Question: why is it that in most games, the FX 8350 seems to compete so well with the 3rd and 4th generation CPUs? I notice that between the 8350 and the 4770K, in most games, there is only a 5-10 FPS difference. So are we saying that an AMD 4 core FX CPU is as powerful as an 8 core hyperthreaded i7?? Please help me...I really don't understand.

Well they aren't a full core, but they're also more than half a core.

kinda like hyper-threading only backwards where the hyper-threaded "core"s don't quite count as a full core.

 

and at what clock speed were the comparisons at and with what GPU, and what game?

[stconquest]

So are we saying that they don't actually have 8 cores?? I was planning on buying an FX 8320, but after reading this article, I am being turned away from AMD ever so slowly. So to start a discussion, we are saying that they only have 4 cores? What about the FX 6000 series?? Do they only have 3 cores?? Someone help me pls. I plan on buying a new CPU and I am now not sure what to buy. I want something that is good for gaming, recording/streaming. As aforementioned, I was planning on buying the 8320, but have now been driven away from them. Another Question: why is it that in most games, the FX 8350 seems to compete so well with the 3rd and 4th generation CPUs? I notice that between the 8350 and the 4770K, in most games, there is only a 5-10 FPS difference. So are we saying that an AMD 4 core FX CPU is as powerful as an 8 core hyperthreaded i7?? Please help me...I really don't understand.

 

Usually single player games are GPU bound.  If you were to run a GTX 980Ti with a CPU, you are more likely to find a scenario where the AMDs start falling off compared to the Intels. 

 

For gaming, the multi-thread support is not as important as the performance on a single core.  Intel is roughly 40% faster in single threaded tasks.  An i5 that only has 4 cores/threads can actually do a good job keeping up with an 8 core/thread AMD CPU because of the faster architecture.

 

The i7 5690X ($1000) wipes the floor with the FX CPUs in every possible task.  The Xeon E3-1230 V3 ($230) beats the FX CPUs in almost every scenario.

 

If you are budget restricted, go with an i5 4460 (best value). 

 

Want the CPU to last a year or two longer than a 4460?  Grab the 4690K so you can OC it when you need the higher clock. 

 

Want multiple thread support along with gaming performance?  Go with a Xeon (E3-1230 V3 or higher)

 

i7 4690K or i7 6700K if budget is not an issue... or even a 5820K.

 

I have literally only seen a single build on here that ever warranted running a FX 8320... that was for a PC that was solely for chess and was tight budgeted.

 

To stream or record you have options like using the iGPU on an Intel CPU or the dGPU itself.

[Guest]

Usually single player games are GPU bound.  If you were to run a GTX 980Ti with a CPU, you are more likely to find a scenario where the AMDs start falling off compared to the Intels. 

 

For gaming, the multi-thread support is not as important as the performance on a single core.  Intel is roughly 40% faster in single threaded tasks.  An i5 that only has 4 cores/threads can actually do a good job keeping up with an 8 core/thread AMD CPU because of the faster architecture.

 

The i7 5690X ($1000) wipes the floor with the FX CPUs in every possible task.  The Xeon E3-1230 V3 ($230) beats the FX CPUs in almost every scenario.

 

If you are budget restricted, go with an i5 4460 (best value). 

 

Want the CPU to last a year or two longer than a 4460?  Grab the 4690K so you can OC it when you need the higher clock. 

 

Want multiple thread support along with gaming performance?  Go with a Xeon (E3-1230 V3 or higher)

 

i7 4690K or i7 6700K if budget is not an issue... or even a 5820K.

 

I have literally only seen a single build on here that ever warranted running a FX 8320... that was for a PC that was solely for chess and was tight budgeted.

 

To stream or record you have options like using the iGPU on an Intel CPU or the dGPU itself.

OK so in my original PC, I have an R9 270X and currently run an A10-6800K (crap, I know). I am a budget gamer and are looking to upgrade to a better CPU. I was originally looking at the FX 8320. I chose the 8320/8350 because it had high core amounts (good for multitasking) and was a good budget CPU. I plan on gaming and recording/streaming. I need help with this because what people don't realise, is that I am a budget gamer. I don't want to pay (for what I consider) overpriced PC parts. I need your help, but I don't want any fanboys and I need legitimate advice

[Guest]

Well they aren't a full core, but they're also more than half a core.

kinda like hyper-threading only backwards where the hyper-threaded "core"s don't quite count as a full core.

 

and at what clock speed were the comparisons at and with what GPU, and what game?

It was the 4770K and the 8350 with two-way SLI GTX 670 (it was a couple of years ago).https://www.youtube.com/watch?v=UUBE9xNz_mQ

Do my question sort of is: due to the latest controversy, should I get an FX 8320 (with 6 cores) or get an FX 6300/6350. I don't want to go Intel Core i7 because the cheapest one (in Australia) with the motherboard and a cooler is like $600

An AMD CPU by comparison is around $400 for the CPU, Motherboard and Cooler. The 8320 is capable of gaming and recording/streaming and should do me well. What should I go with? FX 8000 series (6 cores) or an 6000 series

[KRDucky]

OK so in my original PC, I have an R9 270X and currently run an A10-6800K (crap, I know). I am a budget gamer and are looking to upgrade to a better CPU. I was originally looking at the FX 8320. I chose the 8320/8350 because it had high core amounts (good for multitasking) and was a good budget CPU. I plan on gaming and recording/streaming. I need help with this because what people don't realise, is that I am a budget gamer. I don't want to pay (for what I consider) overpriced PC parts. I need your help, but I don't want any fanboys and I need legitimate advice

I will provide you with my advice coming from me using AMD and Intel hardware for gaming, coding, streaming, 3D modelling, Video editing, Audio production and server roles.

 

You go Intel if you have cash to spend and want the highest IPC you can get for mostly single threaded applications.

You go AMD, if you are on a budget, don't care if you don't have the highest IPC so long as your experience is not degraded. (don't care if it is solid 60FPS so long as it is 30-60 FPS on average). You also go AMD if you are doing multi-threaded applications or applications that can make use of Multi-cores. there is a difference between Multi-threaded and Multi-core. You can be Multi-threaded on a single core that has Hyperthreading making that one core have 2 threads. Server roles that are 64-bit make an excellent example of this. Look at most enterprise grade servers. They are Dual and Quad socket (2 or 4 processors) servers with hundred of Gigabytes of RAM. That is an example of something that makes use of multiple cores. Otherwise, what is the point of more than one core if a program is only going to use one core but 2 threads? These days most games still only make use of at most 2 cores.

 

For the time being, it is more cost effective to stick with AMD. Especially since you will have to buy a new motherboard. When I went from my FX 8120 to my I7-4770k, I had to buy the processor and the motherboard. It ran me around $500 total. A decent change for you would be to swap your A10 for an 860K. Still the same socket but without the APU. it is a decent Quad core gaming CPU using the FM2+ socket. You can also overclock it to hell and back. Stick a decent cooler on it and you are golden. What you might want to wait for is the release of the AMD Zen chips.

 

when streaming, I found my FX chip served me better than my intel chip. the same goes for 3D modelling, and video editing. I used programs that could make use of more than one core.

running Linux is a good way to do that. Under Linux, if I wanted to I could recompile a program specifically for my hardware.

 

FYI, an FX 8320 is 8 cores. Not 6 cores. That is the 6000 series.

 

Also, it is 8 cores. not 8 semi cores equalling 4 full cores. That is Intel math. several years ago CPUs did not have an FPU built in. They were called Co-Processors. And yet, Intel and AMD made CPUs that were considered "full" CPUs. There is no black and white definition of what makes up a "Core". there have been other processors designed in the same manner and they were also considered to be a full processor.

[stconquest]

It was the 4770K and the 8350 with two-way SLI GTX 670 (it was a couple of years ago).https://www.youtube.com/watch?v=UUBE9xNz_mQ

Do my question sort of is: due to the latest controversy, should I get an FX 8320 (with 6 cores) or get an FX 6300/6350. I don't want to go Intel Core i7 because the cheapest one (in Australia) with the motherboard and a cooler is like $600

An AMD CPU by comparison is around $400 for the CPU, Motherboard and Cooler. The 8320 is capable of gaming and recording/streaming and should do me well. What should I go with? FX 8000 series (6 cores) or an 6000 series

 

You do not go with an FX 8XXX if you want to save money.  You do the 4440 and a H81 (stock cooling).

 

PCPartPicker part list / Price breakdown by merchant

CPU: Intel Core i5-4440 3.1GHz Quad-Core Processor  ($179.99 @ SuperBiiz)

Motherboard: MSI H81M-P33 Micro ATX LGA1150 Motherboard  ($42.89 @ OutletPC)  <Note: no Front USB 3.0 header

Total: $222.88

Prices include shipping, taxes, and discounts when available

Generated by PCPartPicker 2015-11-10 00:16 EST-0500

 

or the 4460 and a H97:

 

PCPartPicker part list / Price breakdown by merchant

CPU: Intel Core i5-4460 3.2GHz Quad-Core Processor  ($172.89 @ OutletPC)

Motherboard: ASRock H97M Anniversary Micro ATX LGA1150 Motherboard  ($67.89 @ OutletPC)

Total: $240.78

Prices include shipping, taxes, and discounts when available

Generated by PCPartPicker 2015-11-10 00:17 EST-0500

 

This is the FX 8XXX counter, no cooler so no big overclcock:

 

PCPartPicker part list / Price breakdown by merchant

CPU: AMD FX-8320 3.5GHz 8-Core Processor  ($139.89 @ OutletPC)

Motherboard: Gigabyte GA-990FXA-UD3 R5 ATX AM3+ Motherboard  ($94.99 @ SuperBiiz)

Total: $234.88

Prices include shipping, taxes, and discounts when available

Generated by PCPartPicker 2015-11-10 00:19 EST-0500

 

With Intel, you can use Quicksync and OBS to stream/record video.  It uses the iGPU on the CPU.  ...or you can use the Evolved thingy that AMD has.  Or you can just use the i5 for a guaranteed 720p@30fps stream, as long as your internet connection is good enough.

[KRDucky]

You do not go with an FX 8XXX if you want to save money.  You do the 4440 and a H81 (stock cooling).

 

PCPartPicker part list / Price breakdown by merchant

CPU: Intel Core i5-4440 3.1GHz Quad-Core Processor  ($179.99 @ SuperBiiz)

Motherboard: MSI H81M-P33 Micro ATX LGA1150 Motherboard  ($42.89 @ OutletPC)  <Note: no Front USB 3.0 header

Total: $222.88

Prices include shipping, taxes, and discounts when available

Generated by PCPartPicker 2015-11-10 00:16 EST-0500

 

or the 4460 and a H97:

 

PCPartPicker part list / Price breakdown by merchant

CPU: Intel Core i5-4460 3.2GHz Quad-Core Processor  ($172.89 @ OutletPC)

Motherboard: ASRock H97M Anniversary Micro ATX LGA1150 Motherboard  ($67.89 @ OutletPC)

Total: $240.78

Prices include shipping, taxes, and discounts when available

Generated by PCPartPicker 2015-11-10 00:17 EST-0500

 

This is the FX 8XXX counter, no cooler so no big overclcock:

 

PCPartPicker part list / Price breakdown by merchant

CPU: AMD FX-8320 3.5GHz 8-Core Processor  ($139.89 @ OutletPC)

Motherboard: Gigabyte GA-990FXA-UD3 R5 ATX AM3+ Motherboard  ($94.99 @ SuperBiiz)

Total: $234.88

Prices include shipping, taxes, and discounts when available

Generated by PCPartPicker 2015-11-10 00:19 EST-0500

 

With Intel, you can use Quicksync and OBS to stream/record video.  It uses the iGPU on the CPU.  ...or you can use the Evolved thingy that AMD has.  Or you can just use the i5 for a guaranteed 720p@30fps stream, as long as your internet connection is good enough.

dude, did you not notice that he said Australia?!!!

 

A i5- 4460 in Australia is ~265 AUD. I could not find an I5- 4440 on PCPartpicker's Australia site. The only intel part that was sub 200 AUD is the i3-4340 at 195 AUD.

On the other hand the FX 8320 is 218 AUD with an 89 AUD gigabyte board giving you 307 AUD

 

If we were to use your parts list the first setup (with i3-4340 instead of i5-4440) you would get: 264 AUD

the second would be: 384 AUD

 

And the AMD parts you put: 433 AUD But if would swap the mobo with an Asrock 990FX Extreme3 you get 363 AUD

 

On the other hand, if jkeasley were to keep his board for the A10-6800K he has he could move to the AMD 860K he would spend ~100 AUD from Newegg Australia. If his motherboard was not supporting the AMD Athlon X4 860K he could get the CPU and Motherboard for ~199 AUD with an ASROCK FM2A88X+ Killer motherboard for 99 AUD.

 

So, if he were to go with the 860K, he would get a stronger CPU, and not spend as much as moving to even an Intel core i3 4340 CPU.

 

 

With AMD he can use OBS and OpenCL which is actually better than Quicksync (source: OBS forums) and AMD VCE. Using AMD's VCE you can stream/record with almost no performance loss, at better low bitrate quality then Shadowplay (Reddit:https://www.reddit.com/r/Twitch/comments/2chtng/streaming_recording_with_amds_vce_its_good/ )

 

https://obsproject.com/forum/threads/obs-branch-with-amd-vce-support.13996/.

 

Also, you can NOT guarantee 720p@60fps because it all depends on what he is streaming gameplay of. For example, if he is streaming Arma 3, have fun getting a solid "guaranteed" 720p@60fps. Don't make claims you can not back up because if he goes your route and it backfires, he is going to blame you for telling him something that you can not guarantee.

You sound more like an Intel salesman than someone giving sound, unbiased advice.

[stconquest]

 

dude, did you not notice that he said Australia?!!!

 

A i5- 4460 in Australia is ~265 AUD. I could not find an I5- 4440 on PCPartpicker's Australia site. The only intel part that was sub 200 AUD is the i3-4340 at 195 AUD.

On the other hand the FX 8320 is 218 AUD with an 89 AUD gigabyte board giving you 307 AUD

 

If we were to use your parts list the first setup (with i3-4340 instead of i5-4440) you would get: 264 AUD

the second would be: 384 AUD

 

And the AMD parts you put: 433 AUD But if would swap the mobo with an Asrock 990FX Extreme3 you get 363 AUD

 

On the other hand, if jkeasley were to keep his board for the A10-6800K he has he could move to the AMD 860K he would spend ~100 AUD from Newegg Australia. If his motherboard was not supporting the AMD Athlon X4 860K he could get the CPU and Motherboard for ~199 AUD with an ASROCK FM2A88X+ Killer motherboard for 99 AUD.

 

So, if he were to go with the 860K, he would get a stronger CPU, and not spend as much as moving to even an Intel core i3 4340 CPU.

 

 

With AMD he can use OBS and OpenCL which is actually better than Quicksync (source: OBS forums) and AMD VCE. Using AMD's VCE you can stream/record with almost no performance loss, at better low bitrate quality then Shadowplay (Reddit:https://www.reddit.com/r/Twitch/comments/2chtng/streaming_recording_with_amds_vce_its_good/ )

 

https://obsproject.com/forum/threads/obs-branch-with-amd-vce-support.13996/.

 

 

That Asrock 990 may be okay for a stock FX 8XXX CPU, but don't expect any type of stable OC on it.

 

 

PCPartPicker part list / Price breakdown by merchant

CPU: AMD FX-8320 3.5GHz 8-Core Processor  ($218.00 @ CPL Online)

Motherboard: MSI 970 GAMING ATX AM3+ Motherboard  ($149.00 @ CPL Online)

Total: $367.00

Prices include shipping, taxes, and discounts when available

Generated by PCPartPicker 2015-11-10 16:52 AEDT+1100

 

PCPartPicker part list / Price breakdown by merchant

CPU: Intel Core i5-4460 3.2GHz Quad-Core Processor  ($269.00 @ Centre Com)

Motherboard: ASRock H97M PRO4 Micro ATX LGA1150 Motherboard  ($109.00 @ CPL Online)

Total: $378.00

Prices include shipping, taxes, and discounts when available

Generated by PCPartPicker 2015-11-10 16:53 AEDT+1100

 

 

Intel is still a better option.

[KRDucky]

It was the 4770K and the 8350 with two-way SLI GTX 670 (it was a couple of years ago).https://www.youtube.com/watch?v=UUBE9xNz_mQ

Do my question sort of is: due to the latest controversy, should I get an FX 8320 (with 6 cores) or get an FX 6300/6350. I don't want to go Intel Core i7 because the cheapest one (in Australia) with the motherboard and a cooler is like $600

 

That Asrock 990 may be okay for a stock FX 8XXX CPU, but don't expect any type of stable OC on it.

 

 

PCPartPicker part list / Price breakdown by merchant

CPU: AMD FX-8320 3.5GHz 8-Core Processor  ($218.00 @ CPL Online)

Motherboard: MSI 970 GAMING ATX AM3+ Motherboard  ($149.00 @ CPL Online)

Total: $367.00

Prices include shipping, taxes, and discounts when available

Generated by PCPartPicker 2015-11-10 16:52 AEDT+1100

 

PCPartPicker part list / Price breakdown by merchant

CPU: Intel Core i5-4460 3.2GHz Quad-Core Processor  ($269.00 @ Centre Com)

Motherboard: ASRock H97M PRO4 Micro ATX LGA1150 Motherboard  ($109.00 @ CPL Online)

Total: $378.00

Prices include shipping, taxes, and discounts when available

Generated by PCPartPicker 2015-11-10 16:53 AEDT+1100

 

 

Intel is still a better option.

 

An AMD CPU by comparison is around $400 for the CPU, Motherboard and Cooler. The 8320 is capable of gaming and recording/streaming and should do me well. What should I go with? FX 8000 series (6 cores) or an 6000 series

Actually, the AMd Athlon X4 860K is the better option. He can still OC the hell out of that chip with decent cooling. Have fun dong that with the Intel chip on stock cooling.

[stconquest]

Actually, the AMd Athlon X4 860K is the better option. He can still OC the hell out of that chip with decent cooling. Have fun dong that with the Intel chip on stock cooling.

 

Uhmmm, going with an i5 4460 means that you do not have to overclock.  That's the beauty of the architecture over AMD:  Low power, high performance.

[KRDucky]

Uhmmm, going with an i5 4460 means that you do not have to overclock.  That's the beauty of the architecture over AMD:  Low power, high performance.

ummm. you don't have to overclock an 860K to get great performance. I only brought it up because you mentioned that he should grab the 4690K so that in the future he can OC it for more performance.

 

He wants to save money and still get good performance.

[stconquest]

ummm. you don't have to overclock an 860K to get great performance. I only brought it up because you mentioned that he should grab the 4690K so that in the future he can OC it for more performance.

 

He wants to save money and still get good performance.

 

Is that what I said he should get?  You sure?  :huh:

 

I am done with this tangent.  This is an old argument, I am not going to continue it.

[KRDucky]

Usually single player games are GPU bound.  If you were to run a GTX 980Ti with a CPU, you are more likely to find a scenario where the AMDs start falling off compared to the Intels. 

 

For gaming, the multi-thread support is not as important as the performance on a single core.  Intel is roughly 40% faster in single threaded tasks.  An i5 that only has 4 cores/threads can actually do a good job keeping up with an 8 core/thread AMD CPU because of the faster architecture.

 

The i7 5690X ($1000) wipes the floor with the FX CPUs in every possible task.  The Xeon E3-1230 V3 ($230) beats the FX CPUs in almost every scenario.

 

If you are budget restricted, go with an i5 4460 (best value). 

 

Want the CPU to last a year or two longer than a 4460?  Grab the 4690K so you can OC it when you need the higher clock. 

 

Want multiple thread support along with gaming performance?  Go with a Xeon (E3-1230 V3 or higher)

 

i7 4690K or i7 6700K if budget is not an issue... or even a 5820K.

 

I have literally only seen a single build on here that ever warranted running a FX 8320... that was for a PC that was solely for chess and was tight budgeted.

 

To stream or record you have options like using the iGPU on an Intel CPU or the dGPU itself.

I take it you can't be bothered to remember or read your previous posts.

 

Intel is commonly used for Enthusiast grade builds. Not Budget. That is traditionally AMD's turf.

[CTR640]

So any news? Will AMD lose on this or win? Intel keeps save and yet is still not a damn competition against them.

Damn, I so can't wait to see Intel being dragged to the court for monopoly but that won't happen ever.

I just don't want to see AMD go bankrupt.

[patrickjp93]

So any news? Will AMD lose on this or win? Intel keeps save and yet is still not a damn competition against them.

Damn, I so can't wait to see Intel being dragged to the court for monopoly but that won't happen ever.

I just don't want to see AMD go bankrupt.

Being a monopoly isn't illegal. Abusing your position as market leader or monopoly is illegal. If all your compeition dies off because you're simply the best the market has to offer, you can't be blamed for it.

[CTR640]

Being a monopoly isn't illegal. Abusing your position as market leader or monopoly is illegal. If all your compeition dies off because you're simply the best the market has to offer, you can't be blamed for it.

So if AMD dies off, Intel cannot be taken to the court?