convert to haswell, or stick it out with AMD?

[Rinaun]

Superior cooling increases efficiency. If you aren't yet familiar, review the concept of "voltage leakage". You cite this as an example of AMD's deficiency where I would instead cite it as an example of Intel's penny pinching. Another good example is the TIM on the 4770Ks.

Yes I understand voltage leakage; that's why I suggested he buy a USED old Intel based chip. Intel has penny pinched recently, but I'd still suggest theirs over AMD just throwing voltage hoping for it to stick. You really think they've done 5 year runs at 200W? Add in Overclocking and that sounds like a recipe for failure long-term. I'm also wondering if you have experience with the newer AMD processors like I do. They really heat up a 10x10x9' room in 2 hours of gaming.

no one building a gaming rig seriously considers thermal or electrical efficiency when designing these rigs.

Can you show me any statistics on this? I think you are confusing people who buy titans (a incredibly small fraction of consumers) with people who buy 200$ video cards, like normal consumers. Also you are slightly right in that people don't care about TDP because why would a 14-23 year old living at their parents house care about the electricity bill? For me (and others), I live in CA and since I'm paying extra for my electricity like most in CA, energy IS a concern for me. Just because everyone around you doesn't give a hoot about TDP doesn't mean that's the overall opinion of the market.

What a big shot you are. Please teach me more about overclocking OK? I've been watching everything I can on the utubez and I just can't get those 5 settings right.

This whole conversation degenerated from my point that no one building a gaming rig seriously considers thermal or electrical efficiency when designing these rigs. None of your rant in this paragraph do anything to counter that point. They just make you look like a kid.

So listing my relevant industry experience makes me a big shot? Heh, ok.

Dropping down the 2d mode on a 9800gtx still consumes, what, 40W? That's 40W of totally wasted power, easily equivalent to the inefficiency of the FX-8320.

Yeah see, you have no idea what you are talking about at all and this is why this will be my last reply to you.

I gave the client a free card. My client DOES NOT run their PC 24/7 like I do, thus there isn't any real reason for them to even consider the loss of "40 watts". Secondly, you might want to look up you know....actual statistics like I do.

http://www.guru3d.com/articles_pages/geforce_9800_gtx_sli_review_(bfg),3.html

See that? Do you see the difference in idling between one card, and two? I'll type it out so we all understand you saw it 🙂

9800 GTX 512 MB load system TDP 306 Idle 165
9800 GTX 512 MB SLI load system TDP 419 Idle 168
it's not a gtx+ which is ACTUALLY more energy efficient. So no, your analogy is STILL completely wrong. for 3 watts you get 50% more for free. You can argue until you are blue but yeah have fun because I'm done 🙂.

 

[frozentundra123456]

Power usage is definitely a factor that has to be considered in an accurate estimate of cost of one processor vs another. However much other power one uses is irrelevant to that specific comparison. It just blows my mind that people try to justify totally ignoring it. Your total rig may consume 2000 watts, but if cpu x consumes 100 watts more than cpu y, then that cpu costs more to operate than cpu y. Other power usage is irrelevant for a comparison of the total cost of ownership of the cpus.

It may be a very small amount of money compared to other things, but if one tries to justify one cpu over another on the basis of price, cost of the power used has to be considered.

 

[Rinaun]

Power usage is definitely a factor that has to be considered in an accurate estimate of cost of one processor vs another. However much other power one uses is irrelevant to that specific comparison. It just blows my mind that people try to justify totally ignoring it. Your total rig may consume 2000 watts, but if cpu x consumes 100 watts more than cpu y, then that cpu costs more to operate than cpu y. Other power usage is irrelevant for a comparison of the total cost of ownership of the cpus.

It may be a very small amount of money compared to other things, but if one tries to justify one cpu over another on the basis of price, cost of the power used has to be considered.

Exactly 🙂. Consumers should be aware at how much long-term electricity (and in this case, AC) costs will affect them when comparing processors to purchase. By all means if you live in Siberia and have .03$/kwH, then buy the AMD chip!

 

[TerryMathews]

Power usage is definitely a factor that has to be considered in an accurate estimate of cost of one processor vs another. However much other power one uses is irrelevant to that specific comparison. It just blows my mind that people try to justify totally ignoring it. Your total rig may consume 2000 watts, but if cpu x consumes 100 watts more than cpu y, then that cpu costs more to operate than cpu y. Other power usage is irrelevant for a comparison of the total cost of ownership of the cpus.

It may be a very small amount of money compared to other things, but if one tries to justify one cpu over another on the basis of price, cost of the power used has to be considered.

So to bring this back to my original point, given two choices would you choose a processor that had less total compute performance but better performance per watt or would you elect for the faster processor that was less efficient.

My statement, and I still stand behind it, is that anyone who is an enthusiast would select total performance over electrical efficiency.

 

[TerryMathews]

Can you show me any statistics on this?

Ever been to CPUs and Overclocking? Every +0.01 of Vcore makes the processor less efficient vs stock.

It seems we have a lot of people who post here who aren't overly concerned with that type of efficiency.

Raw performance drives the enthusiast, and I would agree 100% that isn't AMD's forte and hasn't been for a while. Doesn't change the fact that the people here who are talking up the 4770k and 4670k and 2500k aren't doing it on the basis of their thermal efficiency. They're doing it based on the FPS they get as the end result.

 

[nwo]

Depends on the rep you get at Staples, Some won't price match an in store price from MicroCenter.

I tried to get a price match on the 4570 and got shot down. Was told it had to be available online. I also know that Staples did price match the i7-4770 deal from last week.

Well yeah, of course they must have it in stock in order to do a price match. They can't do a price match if they are sold out. Because in order for them to do a price match, you have to place the order and the rep will refund you for the price difference.

I agree, it depends on the rep you get. Same goes with my newegg experience. Some are worse some are better than others.

But the one time I asked staples to do a $149.99 price match for the i5 3570k and they agreed on it, the CPU was in stock (online of course).

My statement, and I still stand behind it, is that anyone who is an enthusiast would select total performance over electrical efficiency.

Heck yeah I would choose total performance over electrical efficiency. My electricity is only about 6 cents per kWh. But I usually go for the best bang for the buck deals.

However, you really have to take both aspects into consideration. If a CPU is less efficient and draws more power, it means it will also produce a lot more heat. This heat is not only gonna make your room super hot in the summer but it will also make you invest a lot more in aftermarket cooling and/or limit your OC potential.

 

[TerryMathews]

This heat is not only gonna make your room super hot in the summer but it will also make you invest a lot more in aftermarket cooling and/or limit your OC potential.

Question is, would that keep you from buying the CPU? Assuming its a good deal up front cooling included.

 

[nwo]

Question is, would that keep you from buying the CPU? Assuming its a good deal up front cooling included.

No, it would not prevent me from buying it if I take cooling and OC limitations out of the picture.

 

[TerryMathews]

No, it would not prevent me from buying it if I take cooling and OC limitations out of the picture.

Which was my point, thanks for confirming.

 

[frozentundra123456]

So to bring this back to my original point, given two choices would you choose a processor that had less total compute performance but better performance per watt or would you elect for the faster processor that was less efficient.

My statement, and I still stand behind it, is that anyone who is an enthusiast would select total performance over electrical efficiency.

We really are not talking about the same thing. The only statement that I am making, is that when one says widget A is a better value than widget B based on price, that the initial cost as well as the cost of ownership must be considered.

 

[sequoia464]

Because in order for them to do a price match, you have to place the order and the rep will refund you for the price difference.

I agree, it depends on the rep you get. Same goes with my newegg experience. Some are worse some are better than others.

But the one time I asked staples to do a $149.99 price match for the i5 3570k and they agreed on it, the CPU was in stock (online of course).

Did I go about this incorrectly then?? Tried to phone order with a price match an i5 4570 from Staples that both Staples and Microcenter had in stock , tried the order over the phone instead of online.

Should I have ordered online and then called Staples for the price match?

 

[Torn Mind]

Passmark is a suitable benchmark only for tasks that tap all cores. Gaming is inconsistent with regards to that.

The speed of something is measured in units per second. What those units are in the context of computers vary, but IPS(instructions per second) is one of those measures. IPS can be derived by multiplying the clock speed by IPC(instructions) per clock. This will get you the IPS of a single core. Speed of multiple cores is the sum of the speed of the single cores.

 

[Rinaun]

Did I go about this incorrectly then?? Tried to phone order with a price match an i5 4570 from Staples that both Staples and Microcenter had in stock , tried the order over the phone instead of online.

Should I have ordered online and then called Staples for the price match?

You should go in person and be SUPER nice. Sounds stupid but I've had the best luck in person being extremely polite. I've had frys price-match some crazy stuff just by being kind 🙂

 

[sequoia464]

You should go in person and be SUPER nice. Sounds stupid but I've had the best luck in person being extremely polite. I've had frys price-match some crazy stuff just by being kind 🙂

I'll try that - thanks. Already have a 4440 but would prefer the 4570. This is an office machine - no overlooking, looking for quiet and stable.

 

[Jacky60]

You didn't read a word i wrote on the last two pages did you?

I read that you enjoy touting 8 cores to what can only be a similarly badly informed audience and you don't know what litecoins are and you think buying AMD CPU's that perform noticeably worse than their Intel counterparts is somehow better 'value' but hey, ignorance is bliss for some peopleD:

 

[Spawne32]

I read that you enjoy touting 8 cores to what can only be a similarly badly informed audience and you don't know what litecoins are and you think buying AMD CPU's that perform noticeably worse than their Intel counterparts is somehow better 'value' but hey, ignorance is bliss for some peopleD:

Go back and read it again.

 

[Jacky60]

You asked people's opinions and you got them, you disregarded most that conflicted with your pre-conceptions. I'm not reading any more of your drivel. 🙂Good luck heating the room over the garage.

 

[inf64]

Looking at the doom and gloom posts in this topic one would wonder if those "super slow" AMD chips can even run Windows OS or games from 2007 properly... It's kinda ridiculous :sneaky:

On a serious note, Spawn have you tried to pinpoint the stability problem? Board's VRM is weakish for 8320?

 

[Spawne32]

Looking at the doom and gloom posts in this topic one would wonder if those "super slow" AMD chips can even run Windows OS or games from 2007 properly... It's kinda ridiculous :sneaky:

On a serious note, Spawn have you tried to pinpoint the stability problem? Board's VRM is weakish for 8320?

Basically yes, I can only run 6 cores to keep the VRM stable. But i am able to run those 6 cores up to 4.4ghz, but 8 even at stock it will throttle due to voltage alone. The only thing I really need is a better mATX motherboard, but it seems that is NEVER going to happen.

 

[USER8000]

Basically yes, I can only run 6 cores to keep the VRM stable. But i am able to run those 6 cores up to 4.4ghz, but 8 even at stock it will throttle due to voltage alone. The only thing I really need is a better mATX motherboard, but it seems that is NEVER going to happen.

I think I would stick with what you have and see what will come in another year,ie,Intel will hopefully have their next gen out by then. Also,you should be able to save some more money up by then too.