AMD FX-4170 vs Intel Sandy Bridge i5

[rEdsKu11z]

Hello i was wondering what cpu would be better for gaming the AMD FX-4170 or the Intel Sandy Bridge i5? i noticed that the i5 is more used in gaming rigs but the FX-4170 seems to have better performance then the i5 (4.2 ghz to 3.6 ghz).

Sorry if this is an easy question but i do not know processors as much as i thought.

Thank you, for you time to read this

Steve

 

[rishiswaz]

So how does a BD Module equal a core then?

A BD is exactly like any other core but it does not have a dedicated 256bit FPU, each core has a single 128bit FMAC FPU and for 256bit floating point integer values a module is used. Now unless you build a machine only for benchmarks, or do large amounts of vector scaling or anything that requires double precision calculation or above a 128bit FMAC is just fine. Most benchmarks however use double precision, and more advanced compute programs use anywhere up to octo precision calculations. If you are doing high precision calculations go with the X79 chipset and get yourself a 3960X, maybe even two and run them in a double array, but for day-to-day activity and gaming FX processors do not experience any PHYSICAL problems. The problem with BD is that it focused on a shift to high core count multi-threaded applications before a majority of programs shifted to being able to use more than 4-6 cores. AMD shot itself in the foot with the FX architecture, The K10 was great and they could have done a die shrink of it and gone a lot farther

 

[rishiswaz]

AMD needed better module/core synergy and better cache speeds and distributions. Also they could learn from intel and go for less pipelines with longer lengths, and give higher speed L1, L2, and get rid of or reduce function of WCC rather than low speed L1 and L2 but some high speed WCC.

 

[cbrunnem]

AMD needed better module/core synergy and better cache speeds and distributions. Also they could learn from intel and go for less pipelines with longer lengths, and give higher speed L1, L2, and get rid of or reduce function of WCC rather than low speed L1 and L2 but some high speed WCC.

man if it was that simple then they would have do it. The AMD engineers arent idiots.

 

[rishiswaz]

man if it was that simple then they would have do it

It is not that simple, they would need to find a way to either shrink the die and include these features, or to keep the current die size and add these among other features. It would be an engineering masterpiece though if AMD got BD core/module implementation down with up to 12 cores/6 modules on a desktop environment, and even more for servers. But we will have to see, and the new architecture structure is going to stay till around 2014, then we might see a K11 or something even more exotic. Who knows, Excavator might bring AMD back out of the grave Bulldozer put it in.

 

[cbrunnem]

man if it was that simple then they would have done it. The AMD engineers arent idiots.

theres an if in there. i was saying exactly what you are without outright saying it. Paper engineering is easy, implementation is not.

It is not that simple, they would need to find a way to either shrink the die and include these features, or to keep the current die size and add these among other features. It would be an engineering masterpiece though if AMD got BD core/module implementation down with up to 12 cores/6 modules on a desktop environment, and even more for servers. But we will have to see, and the new architecture structure is going to stay till around 2014, then we might see a K11 or something even more exotic. Who knows, Excavator might bring AMD back out of the grave Bulldozer put it in.

normally the only way to catch up in situations like this is to throw in a smart way money at the problem but AMD has no money for that type of R&D so idk how they will ever catch up.

 

[rishiswaz]

I think the ATI acquisition was the worst move AMD made for themselves and bit off more than they could chew and now have run into a decision of which to allocate more funding to. Their failing CPUs with 18% market share, or their GPUs with around a 45% market share. Re-selling it now would be a PR nightmare and getting rid of their CPU division would give Intel a 100% market share of consumer level CPUs.

 

[cbrunnem]

I think the ATI acquisition was the worst move AMD made for themselves and bit off more than they could chew and now have run into a decision of which to allocate more funding to. Their failing CPUs with 18% market share, or their GPUs with around a 45% market share. Re-selling it now would be a PR nightmare and getting rid of their CPU division would give Intel a 100% market share of consumer level CPUs.

i dobt that the two entities are combined like you say. the gpu department prolly cant pull a ton of money from the cpu part of it and vise versus. if they are combined i cant see that being smart.

 

[rishiswaz]

They can improve their bottom line by cutting consumer level R&D, focus on low cost APUs allowing them to gain OEM markets, or they could go out to enterprise solutions only with embedded and server processors being their new main focus. AMD solutions power Wii, Xbox 360, Wii U, the next Xbox has rumors of being AMD powered, and same rumors circling around PS4. Also Boeing powers some of their aircraft's cockpits with AMD embedded CPU and GPUs. I think they do not need the consumer market and could drop funding for enthusiast level CPUs after Excavator. This would save them money and also allow for more resources to be given to healthier branches of the company. This is all hypothetical though and AMD obviously has it's reasons for staying in the game.

 

[Proximon]

This topic has been closed by Proximon

 

[Proximon]

Best answer selected by Proximon.