should I abandon the AMD ship? (Updated)

[MiddleOfTheRoad]

I see your point, but I think Anandtech would do some due diligence when compiling their figures. Here's a quote from the article from which the graphs were pulled. I'll leave it to you and other readers to decide if Anandtech should get the benefit of the doubt over your objections:

Edit: Fixed goof in quote, and here is a link to the article:

http://www.anandtech.com/show/8312/state-of-the-part-cpus

Which does prove my point. I don't think I've ever spent more than $50 on a motherboard for an AMD CPU in 15 years of building PC's. Heck, it's not difficult to get one for $30 after rebate from Newegg.

But the motherboard prices are all over the place in this "system build" comparison -- Comparing a $75 FM2 motherboard with a $225 LGA 2011 motherboard is about as Apples To Oranges as it gets.

 

[MiddleOfTheRoad]

Except that Intel's niche could be described as 90%+ of the market.

Imagine that -- a company that is 5 times larger has 4 times the market share.

 

[Enigmoid]

Considering Passmark and Anand posted very similar results for Processor Performance to Price ratios -- I'd say they are pretty spot on.... Not very many reviewers provide price / performance ratios.... But Passmark compiles thousands of user scores to get their data.

Passmark is a good measure of a perfectly threaded test using older instructions. Its okay as a test goes but I wouldn't say its any better than something like geekbench. Thousands of review scores mean nothing if the test is poor (just a point of clarification).

Because there are a host of variables that can skewer the results of a system build. Since you can't mount an AMD chip and Intel chip on the same motherboard -- there is no possible way you get an accurate apples to apples system cost. How do we know the builder didn't use a expensive AM3 board versus a cheap 1150 to paint a better picture of the system value? Plus additional components skewer the results further (Nvidia Cards seem to perform better than Radeons on AMD CPU's -- throwing off the value benchmarks even further depending on how those benchmarks were run and with which components). Synthetic CPU scores remove a ton of (potentially misleading) unknowns that are included in these vague "system builds."

"System Builds" have far too many unknowns to ever be useful for comparisons. How do we know they were running same Ram speeds? Equivalent motherboards? Equivalent motherboard prices? It just gets too unclear... Too many wild cards....

This is true but the component costs are provided and are quite moderate.

f you purchase a motherboard, RAM, SSD, case, and PSU it will tack on $350-$510. I used the base set of components from the recent Budget PC Guide (minus the HDD), and then chose reasonable motherboards for the various platforms ($75 for FM2/FM2+, $80 for AM3+, $100 for LGA1155, $90 for LGA1150, and $225 for LGA2011 with $10 extra on RAM to get 4x2GB instead of 2x4GB). You could certainly spend a lot more on system components, but while less expensive motherboards exist for some of the platforms, I'm a bit wary of any motherboard priced under $70. The result is that the minimum cost for an entire system ends up being around $425-$500 for the less expensive CPUs/APUs, and $600 for something like the i5-4690K, $670-$705 for Core i7 LGA1150 systems, and as much as $1559 for the i7-4960X. Using the complete system prices (including a $50 discrete GPU on the AM3+ and LGA2011 systems), the performance to price charts look as follows:

You can certainly cut the fat here and there but its a fairly good approximation of a balanced build.

Well, I call complete B.S. -- you are such a fanboy........

Just linked the chart you clearly loved ignoring (which was directly above the one you just cherrypicked from Anand's web site) -- This one focusing on the CPU itself...... which BTW is very similar to Passmark's ratings.

Backing up exactly what I just stated. Unless it is a Pentium or Celeron, Intel CPU's generally have pretty lame bang for the buck. If bang for the buck is the ultimate goal for the build -- AMD still remains the better option as soon as the CPU price exceeds a Pentium/Celeron.

Looking at the three primary charts, none can be dismissed.

Sometimes absolute performance trumps any measure of bang for the buck. If you are making money using the computer and performance is a bottleneck it makes not sense not not buy the best parts available. For instance for someone like an engineer, if a K6000 vs. the K5000 boost productivity by even 5% then for an engineer being paid $100,000/year, the card will have paid for itself in about 8 months. Software licences, maintenance and support, cooling/space/power, etc drive up the importance of this chart.

While somewhat relevant, this is probably the least useful chart. A processor is useless without a system to run in. The direct values on this chart only make sense when a person has a complete system with a compatible motherboard. However, this chart is very useful as it helps pick the most efficient product within the price range you are willing to spend. If you tracked performance/system cost the balance would be toward high end components at high end prices. By limiting system price or requiring a given amount of performance, the performance/component is a much better way to measure value. It is important to be flexible however and make according adjustments.

Very important to industry and business where the amount of performance is important. Good example would be this forum's servers. The total server price is the important thing and due to the large amount of budget required for many server setups this is the factor that is more important. An increase in performance for 10% more cost while the system performance increases 20% is great bang for the buck meaning you either have more throughput for a slightly higher cost or you can reduce costs by cutting out entire servers. System costs can dwarf the CPU costs (if you need a ton of RAM). This is going to include the total lifetime cost of the system in this case. Systems such as laptops rely exclusively on this metric. Building on a budget you get the best performance you can at a given price (with the desired quality) and ultimately this is the metric you will use to evaluate the build.

Since tom's is being quoted more

http://www.tomshardware.com/reviews/build-mainstream-enthusiast-pc,3944-15.html

That is how you are going to compare a completed given build.

For consumers in this forum, all three charts matter though the amount of bias toward a particular selection criteria will depend on the user. Absolute performance is a must (why so many people are using high end poor price/perf graphics cards). On a budget the component cost/price is important while for a new build, performance/system price (at a given budget) is the driving factor. It makes little sense to stick to a strict $800 budget if you can add 35% GPU performance by spending $40 more. A compromise must be made, you can't pick one chart and leave the other there.

 

[Ramses]

Nobody, least not me, is saying they don't use more power (in fact I did the math and posted what it costs me), or that the Intel stuff isn't better in some, games. I still have a bucket full of reasons that I'm happy as a clam with an FX system and detailed reasons why, and nobody has yet to show me a reason to "upgrade". Less power usage isn't enough, better performance in games (that I don't play) isn't enough. What else ya got? Really I like upgrading and building computers, give me a good solid reason to buy an I5 or I7. Tell me what I'm missing out on in the 8-15 hours I spend in front of a computer for work and play every day that I'm not aware of after fifteen+ years of building and using computers daily? I don't see it. Intel makes fine CPU's, better in a lot of measurable ways than AMD and I'm glad they do. When the baseline is so damn high already though, it's hard to get excited over something that's further better than something that is already really really good. When you can show me a 25% across the board improvement for the same money, I'll start getting ready to jump, whoever makes it. I got no loyalty to anyone but myself.

To reiterate..

We all know the FX, and it's chipset sucks a lot relative of power.
I can handle $20 a month or so. It's BS living money imo, if I didn't spend it on power I'd spend it on some other crap. Don't care till it's about $50-$60 a month, even then it's just worth thinking about.


We all know the single core and gaming performance is quite a bit less than Intel stuff.
(Though I have yet to be prevented from playing any of the forty games in my Steam account maxed out at 1080P just peachy, I don't play a lot and I don't play online, and I went in fully expecting to lower some settings on newer games when they get cheap enough that I'm willing to buy them. If I wasn't willing to do that I'd have bought Intel)

We all know that in the higher end FX, the price difference isn't all that significant between high-end AMD and mid-range Intel.
I don't know if you guys have ever read it but in Paradise Lost there was that line "better to rule in hell than serve in heaven". Being that the performance deficit is so low for me personally, I'd much rather have a high-end best of the breed AMD chip than a common midrange Intel chip.
You either understand that or you do not, it's a personality based thing and is neither right nor wrong, it just is.

So..
None of those deficiencies, in the slightest, prevent me from thoroughly enjoying this FX based box I have now.

Let me repeat..

None of those deficiencies, in the slightest, prevent me from thoroughly enjoying this FX based box I have now.

At the risk of being vulgar, a 9590 with 16gig of cas7 running two 280x's off a 500gig 540evo is fast as a motherf%^$%r.
I want for nothing, I wait for nothing, other than my pipe to the net which is above average itself. The power usage is functionally irrelevant, so is, apparently, the lackadaisical single core performance. Really, I love upgrades, if I had a good excuse to go buy an I5 or I7 setup, I would. I'd spend the 80 hours, that isn't an exaggeration, to pick the very best motherboard out of the freakin 200 there probably are now, and I'd find the very best ram and have it in a couple weeks. There is just zero motivation to do it. Maybe in a year, or two years, not now. Deficient or not this FX box performs perfectly for a moderately above normal user like myself and I fully expect it to keep doing so for several years at least.
There is no terrible, there is not flat-face-falling, there is nothing but a really nice, fast computer. That's it. It wasn't especially cheap, it doesn't get especially hot. Intel has some faster stuff. Maybe next year AMD will have some faster stuff, maybe not. Right now and for the foreseeable future this was a great box to build and is a blast to use.




As for the system stuff, I get that you want the board/chipset power factored into the equation, I'd like that too, but it just isn't possible. Too much variation, too many boards, and it only gets worse when you go out from there. I encourage you to try and compile the data though if you think it's possible, I'd start with 990FX since there are way fewer of them before tackling the Intel stuff.




It takes a big person of a significant character to say, own a Ferrari and still smile and be happy for and interact and encourage a guy with say, a Mustang. Sure the Ferrari is better(just go with the analogy), but the Mustang is mighty damn good and plenty capable itself.
It's a social skill.
I'm comfortable and a big enough person to admit that my CPU/Chipset is deficient in some areas and that others are faster. I'm OK with that, it don't bother me a bit because I bought it of my own free will after researching and discovering all the stuff above, with my own money and I'm even happy with my decision, after swapping boards, psu's, cases and video cards till I found a good combo.
Believe me, the board and CPU were a small price compared to the rest of this thing, maybe a quarter of the cost.
Anyway, I'm OK with this and I'm happy for a guy with an I5 or an I7.
If a person, like just happened the other day on here via PM, asks me my opinion on my rig and AMD vs Intel, I tell them pretty much what I typed above and let them make there own decision like a grownup does.
I expect that sort of performance from other people, and I don't understand when it isn't forthcoming, so I keep repeating myself. I'm going to have to work on that.

I would even argue that the overall system cost is the true measure of the value of a component. As far as I know, no one has bought a cpu and set it on their desk to look at it, you install it in a system to perform a task. And those who are making excuses for AMD in these threads seem to conveniently forget that some of the initial cost savings will be lost due to increased power consumption. We just went through this "discussion" ad naseum in another thread, and no matter how much one wants to minimize it, it is a factor.

It is also not true that the cpu doesnt matter for gaming. It does depend on the game and the settings, but in cases like this World of Tanks a 4670k is 60% faster than an 8350 at very common gaming settings (single 780Ti, 1080p, one of the most popular on-line games). In fact this game illustrates the big problem with the 8350-- lack of well rounded gaming performance. In some games it performs well, but in anything that is not highly threaded and demands good single core performance, it falls flat on its face, in contrast to an i5 or i7 which performs well in both scenarios.

 

[KingFatty]

I still don't understand the system charts. I mean, assume you pick two sets of identically priced components, the ram, mobo, psu, etc. etc., and stick AMD or Intel chip in it, wouldn't the total system performance just be a reflection/factor of of the processor performance chart?

And then considering price, wouldn't the total system performance per price again be a reflection? I don't see how the different system values can fluctuate so much, assuming the non-processor components are similar (ram etc)? It never made sense to me???

 

[Enigmoid]

I still don't understand the system charts. I mean, assume you pick two sets of identically priced components, the ram, mobo, psu, etc. etc., and stick AMD or Intel chip in it, wouldn't the total system performance just be a reflection/factor of of the processor performance chart?

And then considering price, wouldn't the total system performance per price again be a reflection? I don't see how the different system values can fluctuate so much, assuming the non-processor components are similar (ram etc)? It never made sense to me???

The motherboard prices are different (which is being done in that comparison). Think of it as a big statistical weight towards faster components.

For instance CPU A performs 50% better than CPU B at 200% the cost. CPU B retails at $100.

perf/cpu price

A= 1.5/2 = 0.75
B = 1/1 = 1.00

B gives better CPU perf/$.

However the system costs are around $500 (arbitrary number here).

perf/system price

A= 1.5/7 = 0.21
B = 1/6 = 0.17

B gives better system perf/$.

The usefulness of this depends on your budget and the aim of the build but system perf/$ will tend toward better performing more expensive components. Essentially its a measure of all the money you would spend on the build relative the the performance you would get out of that build.

(The example above could be thought of as an i3 vs. i5 build).

 

[crashtech]

Figuring in the whole cost of the system tends to have a leveling effect because the other components tend to be closer in price. The CPU represents a fairly small percentage of the total cash layout.

 

[MiddleOfTheRoad]

I still don't understand the system charts. I mean, assume you pick two sets of identically priced components, the ram, mobo, psu, etc. etc., and stick AMD or Intel chip in it, wouldn't the total system performance just be a reflection/factor of of the processor performance chart?

And then considering price, wouldn't the total system performance per price again be a reflection? I don't see how the different system values can fluctuate so much, assuming the non-processor components are similar (ram etc)? It never made sense to me???

That's more or less the point I was trying to make -- but you definitely articulated it better.

If every other component is equally priced, why do the system build charts vary so wildly from the synthetic cpu comparison? Simply because the other components in the system builds were not equally priced..... Thus making the whole system build benchmarks murky and pretty tenuous.

If they really wanted to make the system build benchmarks useful -- they should have purchased equally priced motherboards and other components. Otherwise, the baseline is pretty much garbage. It really was a dumb comparison because the motherboards were so dissimilar as stated.

 

[MiddleOfTheRoad]

The motherboard prices are different (which is being done in that comparison). Think of it as a big statistical weight towards faster components.

For instance CPU A performs 50% better than CPU B at 200% the cost. CPU B retails at $100.

perf/cpu price

A= 1.5/2 = 0.75
B = 1/1 = 1.00

B gives better CPU perf/$.

However the system costs are around $500 (arbitrary number here).

perf/system price

A= 1.5/7 = 0.21
B = 1/6 = 0.17

B gives better system perf/$.

Not really -- because the system build still seems rigged by Anand's original comparison.

It is easy to buy identical RAM, Hard Drive/Solid State, Video Card, Case and Power Supply and use identical priced components in both.

Really the only variable is the CPU price and motherboard -- and I can find AM3/FM2 motherboards for an identical price to an 1150 motherboard..... So really the only cost difference is between the CPU's. This "System Build" benchmark seems like a thin attempt at buying a more expensive AMD motherboard (than the equivalent Intel motherboard) to cook the Intel build into appearing like a better value. The Processor Performance / Retail Price benchmark is just a heck of a lot more straightforward.

Even the electricity argument is skating on thin ice for many of these reviews. Because an AMD FX-8350 running a GTX 980 is going to use a ton less power than a Core i7 with a Radeon R9 295X2. The CPU ain't the only thing pulling the juice and many people gloss over the GPU in the majority of these arguments.

I'm never met someone in my life that built an AMD system for the same price as an equivalent Intel setup. People use AMD video cards or CPU's because their goal is nearly always to do a much cheaper build. AMD does generally give you more hardware for less money.... But you usually sacrifice software drivers, heat and power consumption as a result.

 

[Subyman]

We all know the FX, and it's chipset sucks a lot relative of power.
I can handle $20 a month or so. It's BS living money imo, if I didn't spend it on power I'd spend it on some other crap. Don't care till it's about $50-$60 a month, even then it's just worth thinking about.

What about the 198lbs of coal that was burnt for that month? Its crazy when you start to think of the resources involved with our energy usage. People don't understand power efficiency in computers. "I want more compute, I don't care about how many watts it takes!" Until you realize how many lbs of coal are consumed. It is roughly 1.1lbs per kwh.

 

[Ramses]

What about the 198lbs of coal that was burnt for that month? Its crazy when you start to think of the resources involved with our energy usage. People don't understand power efficiency in computers. "I want more compute, I don't care about how many watts it takes!" Until you realize how many lbs of coal are consumed. It is roughly 1.1lbs per kwh.

I consider it a benefit of living with the danger of a nuclear power plant a stones throw from my home. Further, I've been living cheap and small and driving a vehicle that get's 42mpg for the last sixteen years. I have some carbon credit to burn. I'm in Cali, way more concerned with water usage than power. Though I guess an FX might dehumidify the room more... :sneaky:

 

[Ranulf]

What about the 198lbs of coal that was burnt for that month? Its crazy when you start to think of the resources involved with our energy usage. People don't understand power efficiency in computers. "I want more compute, I don't care about how many watts it takes!" Until you realize how many lbs of coal are consumed. It is roughly 1.1lbs per kwh.

Wait till the greenies make it so we only have wind and solar. There are idiots here in the northwest who want to stop hydro.

I got a good laugh when running the numbers on the i3 T chips earlier this year. $10 more and it would take me 2.5 years to make that back based on my power rates of $.12/kwh. All for a 15% slower chip (4330 I think it was).

 

[AtenRa]

but in cases like this World of Tanks a 4670k is 60% faster than an 8350 at very common gaming settings (single 780Ti, 1080p, one of the most popular on-line games).

According to Steam, World Of Tanks is not even registered in "Top games by current player count", unlike War Thunder.

Screenshot made today at 9:45am Local Greece time.

http://store.steampowered.com/stats/

On the other hand, it is in 13th place in Raptr's August "Most Played Games"

http://caas.raptr.com/most-played-p...-legends-world-of-tanks-get-an-esports-boost/

Well i wouldn't say it is one of the most popular on-line games, if you combine Steam and Raptr data base it would not be even in the Top 20.

 

[Seba]

According to Steam, World Of Tanks is not even registered in "Top games by current player count", unlike War Thunder.

World Of Tanks is not on Steam. So it will not show up in any Steam ranking.

 

[Ramses]

Is that the one with the decade old game engine that can't do multicore?
Generally what happens with stuff like that is if I can't tweak it and get it to run ok on whatever I'm running, Intel or AMD, I just move on to the next game. The processing HP is there, if devs can't write to use it, tuff nookie imo. Though I'd be willing to bet it's plenty playable on my box from the couple benches I saw in google.

 

[AtenRa]

World Of Tanks is not on Steam. So it will not show up in any Steam ranking.

Hell i just realized its not on Steam. Thanks for pointing out.

But again, combining both Steam and Raptr games it will strangle to be on TOP 20.

 

[Dribble]

Is that the one with the decade old game engine that can't do multicore?
Generally what happens with stuff like that is if I can't tweak it and get it to run ok on whatever I'm running, Intel or AMD, I just move on to the next game. The processing HP is there, if devs can't write to use it, tuff nookie imo. Though I'd be willing to bet it's plenty playable on my box from the couple benches I saw in google.

So essentially what you are saying is buy AMD and boycot games that aren't well enough optimized enough for your low IPC processor. Alternatively you could just buy Intel and play them all fine. WOT is a huge game, 100 million+ players, probably several million play a day. I don't want my processor choice to get in the way of playing games, I just want to play whatever game I fancy.

 

[AtenRa]

I don't want my processor choice to get in the way of playing games, I just want to play whatever game I fancy.

You can play WOT even with an APU like Kaveri, I am sure the FX + dGPU will not get in the way

 

[Maximilian]

I think we all know AMD chips can play games... they just do it worse than Intel, that's the notable point.

 

[Enigmoid]

Not really -- because the system build still seems rigged by Anand's original comparison.

It is easy to buy identical RAM, Hard Drive/Solid State, Video Card, Case and Power Supply and use identical priced components in both.

Really the only variable is the CPU price and motherboard -- and I can find AM3/FM2 motherboards for an identical price to an 1150 motherboard..... So really the only cost difference is between the CPU's. This "System Build" benchmark seems like a thin attempt at buying a more expensive AMD motherboard (than the equivalent Intel motherboard) to cook the Intel build into appearing like a better value. The Processor Performance / Retail Price benchmark is just a heck of a lot more straightforward.

Yes you can buy cheap $40 FM2+ LGA 1150 motherboards.

($75 for FM2/FM2+, $80 for AM3+, $100 for LGA1155, $90 for LGA1150, and $225 for LGA2011 with $10 extra on RAM to get 4x2GB instead of 2x4GB)

This is a very reasonable approximation however. I don't see this as an attempt to buy a more expensive AMD motherboard at all. Even if you drop $25 from the AMD motherboards it isn't going to affect the system value by more than 5-10%.

 

[crashtech]

Which does prove my point. I don't think I've ever spent more than $50 on a motherboard for an AMD CPU in 15 years of building PC's. Heck, it's not difficult to get one for $30 after rebate from Newegg.

But the motherboard prices are all over the place in this "system build" comparison -- Comparing a $75 FM2 motherboard with a $225 LGA 2011 motherboard is about as Apples To Oranges as it gets.

I don't get the objections to the methodology. A $50 FM2+ is unlikely to have the same feature set as a $90 LGA 1150 board, so to set a $75 price point seems fair to me. This still gives an advantage to AMD in the calculation. Also I don't get the objection to the $225 X99 board. I have seen you tout the availability of $200 X99 boards, the extra $25 tacked on gives Intel a disadvantage in the calculation, not an advantage. If there is true bias in the Anadtech comparison, I just don't see it, but believe that I want to see it if it is really there.

 

[AtenRa]

Well if you asked me i wouldnt use $70-80 boards for budget build systems with AMD A6-6400K and Intel Celeron G1850.

There are excellent FM2+ boards starting at $50 with or without AR and Intel 1150 B85 chipset boards starting at $45,99 having almost the same features.
You can even go lower to A58 and H81 chipsets.

But you get more features with $229 Socket 2011-3 boards and that is something that cannot been seen in the performance per price graphs.

 

[davie jambo]

I think we all know AMD chips can play games... they just do it worse than Intel, that's the notable point.

Agreed

I would also state that the difference is not that much between them , not as much as people on the internet seem to think anyway

 

[MiddleOfTheRoad]

I don't get the objections to the methodology. A $50 FM2+ is unlikely to have the same feature set as a $90 LGA 1150 board, so to set a $75 price point seems fair to me.

The purpose of a benchmark is to measure the same task on competing hardware. If the specs of the other components in these system builds are all wildly dissimilar why even bother running a benchmark? You could simply fall back to the old adage that "You Get What You Pay For" and not waste the time on benchmarking.

Accurate benchmarks rely on a good baseline for comparison purposes. These system builds fail immediately because a consistent baseline was never established. Apples to Oranges -- which is better? Exactly, totally murky.

 

[Ramses]

So essentially what you are saying is buy AMD and boycot games that aren't well enough optimized enough for your low IPC processor. Alternatively you could just buy Intel and play them all fine. WOT is a huge game, 100 million+ players, probably several million play a day. I don't want my processor choice to get in the way of playing games, I just want to play whatever game I fancy.

Not at all and this is an example of a communicative fallacy.
I'm saying that I bought an AMD chip knowing full well it's strengths and weaknesses as any smart consumer should in this day and age, and enjoy having the luxury of choosing from the plethora of other games that are better written and do run well on lots of slower threads rather than one or a few stronger ones. Nor am I advocating anyone else do that unless they want to, and clearly lots of folks do want to and that's fine too. In this specific case, I suspect I could play that game just fine looking at the AMD benchmarks on a thread from the forums, additionally, I went looking a few months ago for a WWII tank sim of some sort, on steam since I rarely use anything else for gaming, and found several. I'm sure it's a fine game but it's a big world these days. It's a matter of principle, and principle is by it's very nature a personal thing. If one needs an Intel chip to play a certain game, I'm happy they are available. I've had them before and will probably have them again.

This is another one of those things that falls under the "let me reiterate" category.

If gaming is really really important, one should buy Intel right now.


We all know this.

Clearly one can game a plenty on AMD stuff, but we also all know that some games cannot or will not effectively use multiple cores. This is no different in spirit imo, at all, than some games playing better on ATI vs Nvidia cards, and it has been this way since the days of the original Gforce 256. I actively avoid GPU-branded games for the same principle based reasons. I vote with my wallet. And that practice I do advocate. :thumbsup: