Best CPU for a budget gaming PC: Athlon X4 860K vs. Pentium G3258

[Steve]

I can totally appreciate the amount of work you put into this article, testing 21 games with this many different configurations is no easy task and I'm sure it took a considerable amount of time. That being said, while it is totally possible, and even likely that your intentions were not malicious in writing this report, it is still hard to come to the conclusion that you did without seeing all the data. That is all I'm saying.

You make it sound like you know the minimum frame rates or frame time data would have favoured AMD which isn’t true. I have already explained why we didn’t include a mountain of data.

If you have evidence that shows our results to be inaccurate by all means serve it up. Short of doing that I see no need to continue this discussion.

 

[Guest]

The lower performance and 50% usage on the AMD cpu might be attributed to 4 ALU / 2 FPU design. Programs that use integer math, like gcc fr compiling software, will use all "4 core", as technically the "cores" are the ALUs. Other software can be heavily FPU dependant, lots of floating point math, and that is really where AMD made a poor choice with Bulldozer by only having one smaller sized FPU per pair of ALUs. In programs with lots of floating point math, the AMDs have half the hardware to do he job.

 

[Donny Stanley]

You make it sound like you know the minimum frame rates or frame time data would have favoured AMD which isn’t true. I have already explained why we didn’t include a mountain of data.

If you have evidence that shows our results to be inaccurate by all means serve it up. Short of doing that I see no need to continue this discussion.

My colleague just finished a review of the 7650K APU, which is a similar CPU to the 860K and compared it against the i5 4430 and Pentium G3258 with both the IGP and a dedicated GPU (R9 290) the results are very different than what you post here http://www.technologyx.com/featured/amd-a8-7650k-apu-review-the-little-apu-that-could/5/

While we'll will have to source an 860K from AMD and do a proper comparison, I think this definitely shows that not displaying all the data does indeed paint a much different picture from the truth, as the Pentium even with a clock speed advantage gets decimated in most situations when it comes to frame time variance, sure it does win often in average FPS but the issue is when you get drops as low as 4 FPS, you'll have a lot of stutter and a horrible experience.

 

[Steve]

My colleague just finished a review of the 7650K APU, which is a similar CPU to the 860K and compared it against the i5 4430 and Pentium G3258 with both the IGP and a dedicated GPU (R9 290) the results are very different than what you post here
While we'll will have to source an 860K from AMD and do a proper comparison, I think this definitely shows that not displaying all the data does indeed paint a much different picture from the truth, as the Pentium even with a clock speed advantage gets decimated in most situations when it comes to frame time variance, sure it does win often in average FPS but the issue is when you get drops as low as 4 FPS, you'll have a lot of stutter and a horrible experience.

Those results contradict my own even when looking at the averages. The GTAV result is terrible, certainly not what I found. I also didn't witness any frame rate lag/stuttering in any of the tests with the Pentium G3258.

Those tests were also conducted with a significantly more powerful GPU.

 

[Alex Stevens]

My colleague just finished a review of the 7650K APU, which is a similar CPU to the 860K and compared it against the i5 4430 and Pentium G3258 with both the IGP and a dedicated GPU (R9 290) the results are very different than what you post here

While we'll will have to source an 860K from AMD and do a proper comparison, I think this definitely shows that not displaying all the data does indeed paint a much different picture from the truth, as the Pentium even with a clock speed advantage gets decimated in most situations when it comes to frame time variance, sure it does win often in average FPS but the issue is when you get drops as low as 4 FPS, you'll have a lot of stutter and a horrible experience.

Those results contradict my own even when looking at the averages. The GTAV result is terrible, certainly not what I found. I also didn't witness any frame rate lag/stuttering in any of the tests with the Pentium G3258.

Those tests were also conducted with a significantly more powerful GPU.

Hi Steve, I'm the writer of said article, I do find the fact that our data is so different to be troubling. I wanted to do a comparison to the G3258 and the Athlon anyways, I think your choices on GPUs are also fantastic. One of the other writers on the site has a GTX 960 which I may borrow for purposes of this article as well as working on grabing a 285, although at this point a 380 would be more likely.

Personally I found the stutter, lag and hitching to be a bit much for me in the games we've tested, although you did test more, and you did test some older games as well. I will recommend, as one reviewer to another if you feel there is too much data, omit some games. Finding where frame drops occur is one of the most important things when gauging the quality of a CPU for gaming. I just find it a bit odd that EVERY game here beat the 860K, and not one time did the G3258 lose.

 

[Steve]

Hi Steve, I'm the writer of said article, I do find the fact that our data is so different to be troubling. I wanted to do a comparison to the G3258 and the Athlon anyways, I think your choices on GPUs are also fantastic. One of the other writers on the site has a GTX 960 which I may borrow for purposes of this article as well as working on grabing a 285, although at this point a 380 would be more likely.

Personally I found the stutter, lag and hitching to be a bit much for me in the games we've tested, although you did test more, and you did test some older games as well. I will recommend, as one reviewer to another if you feel there is too much data, omit some games. Finding where frame drops occur is one of the most important things when gauging the quality of a CPU for gaming. I just find it a bit odd that EVERY game here beat the 860K, and not one time did the G3258 lose.

Firstly thanks for being a lot more professional than Donny Stanley, never before have I had a fellow reviewer/editor bash my work and then shamelessly link to their own.

There are plenty of G3258 owners out there that claim they don’t suffer from stutter/lag in the latest games and I certainly didn’t witness any of this either.

I am not sure why you find it odd that when a clock-to-clock comparison is made the G3258 trumps the 860K in EVERY test. The 860K really only has an advantage when it’s clocked 500MHz higher.

Check out Anand’s results published earlier this month….

http://www.anandtech.com/show/9307/the-kaveri-refresh-godavari-review-testing-amds-a10-7870k/6

In their results the Core i5-4690 is 26% faster than the non-overclocked G3258 which works out to be 7fps.

Meanwhile you show the Core i5-4430 to be 110% faster than a heavily overclocked G3258 at the same resolution :S

My results certainly appear to be more in line with theirs, my frames are higher but the performance trends are the same. I likely tested with slightly different (lower) quality settings.

Meanwhile in virtually every game they tested the non-overclocked 3.2GHz G3258 was faster than the A10-7870K which is clocked slightly higher than the 860K.

Anyway I am certainly not going to pick apart your review, I appreciate the work that went into it despite the fact that your conclusion strongly contradicts my thoughts.

Also while we are giving each other constructive criticism, next time look at including some power consumption results as they are a big part of the picture now.

 

[Guest]

Hi im read sometimes this web and I was time thinking about change mi old athlon x3 because mi new game (FIFA 15) was sometimes touching the sky whit about 5-10 seconds at 100% usage

this was terrible I think and reading reviews like that I buy a pentium g 3250 basically a 3258 whitout unlocked multi


but the experience is the same and worse !! playin is far much beter for 10 -15 20 minutes but then a lag thhe game freeze about 10 seconds

my friends whit a phenom II play smoothly than me

athlon II 455

Pentium g3250

at the peak it freeeze l

thanks

 

[Steve]

Hi im read sometimes this web and I was time thinking about change mi old athlon x3 because mi new game (FIFA 15) was sometimes touching the sky whit about 5-10 seconds at 100% usage

this was terrible I think and reading reviews like that I buy a pentium g 3250 basically a 3258 whitout unlocked multi

but the experience is the same and worse !! playin is far much beter for 10 -15 20 minutes but then a lag thhe game freeze about 10 seconds

my friends whit a phenom II play smoothly than me

athlon II 455

Pentium g3250

at the peak it freeeze l

thanks

How much system memory do you have? It looks like 6GB but I just wanted to make sure.

This processor is pretty weak at 3.2GHz, you really need the unlocked model so you can push it past 4.0GHz.

Here is a video of the G3258 @ 4.4GHz playing FIFA 15...

Frame rates look solid without lag.

 

[Guest]

I have installed 6gb kingston hyper x fury 4g + 2gb kingston value (I try both separately)
http://www.kingston.com/latam/hyperx/memory/fury

defautl clocked at 1333 because cpu dosnt support quicker ones

http://ark.intel.com/es/products/83538/Intel-Pentium-Processor-G3250-3M-Cache-3_20-GHz


is not about cpu performance its about core count a phenom II (teorically slower than the 860k) run smoother

in this case the amd is better choice because the real people dont OC only experts and people who is interested

if I sale a cpu dont do OC on it because if fails dont have chance to use the warranty

another point its that I need a heatsink whit a prize of half the cpu to OC to 4.4 and a motherboard z97 touching the amd fx or i3 prices
I can buy a x4 860k and A88 motherboar 20-30 $ cheaper its about 1/4 the price o the sistem

 

[Guest]

Well-written, detailed article. However, you should also have used as a '3rd testing scenario' the Mantle option for Battlefield 4, Battlefield Hardline, Civilization: Beyond Earth, Dragon Age: Inquisition & Thief. The 860k rig would have pulled ahead in all of those titles.

 

[dirtyferret]

Awesome review and really goes to show that outside of brand loyalty there is no reason to choose AMD over Intel for a gaming PC, especially when you look at upgrade paths. Hopefully AMD moves away from the marketing gimmick of more cores for your dollar and one to more efficient cores for your dollar. Lack of competition is hurting all consumers and I for one would welcome the change and start building with AMD CPUs again.

 

[Donny Stanley]

Firstly thanks for being a lot more professional than Donny Stanley, never before have I had a fellow reviewer/editor bash my work and then shamelessly link to their own.

Firstly, let's be clear. I criticized your lack of posting important data, which details the actual user experience. I did not criticize your work because it drew a conclusion which differed from my own views, I was critical because it declared one product as the clear winner without actually providing the whole story.

Second, I never posted my own work, I posted someone else's I even stated such. If you've noticed, my name is not on the article. Also, I had no intentions of posting any sort of links or even referring to with which firm I worked, however you asked for data which supported my claims so I provided it.

In the future, when someone offers you criticism you should probably take a step back and take a deep breath, before getting so defensive.

 

[dirtyferret]

Firstly, let's be clear. I criticized your lack of posting important data, which details the actual user experience.

Who are you to decide which data is important? That is the web site editor's job.

I did not criticize your work because it drew a conclusion which differed from my own views, I was critical because it declared one product as the clear winner without actually providing the whole story.

..and then posted a link to an "article" that declared a winner without actually providing the whole story, the difference being the winner is the brand you root for.

 

[Donny Stanley]

Who are you to decide which data is important? That is the web site editor's job.

Well, that's pretty easy. I'm a reader. I may be a hardware reviewer, but I still read reviews from other outlets all the times, after all we don't get all the hardware I'm interested in so I still get a lot of information from other sites. That being said, as a reader when I look at an article such as this one that only displays average FPS when I know for a fact that you can have higher averages and still deliver a poorer experience, I feel like I'm not being told the whole story.

[QUOTE="dirtyferret, post: 1482970, member: 366636"
..and then posted a link to an "article" that declared a winner without actually providing the whole story, the difference being the winner is the brand you root for.
[/QUOTE]

When does Alex ever declare a winner? He stated in his review that it (the 7650K) was a top value, and recommend it for users on a budget, which it is in fact a great option for. The data he posted showed it beat the G3258 in a variety of tests, and he did not exclude data to make one product look better than another.

 

[dirtyferret]

Who are you to decide which data is important? That is the web site editor's job.

Well, that's pretty easy. I'm a reader. I may be a hardware reviewer, but I still read reviews from other outlets all the times, after all we don't get all the hardware I'm interested in so I still get a lot of information from other sites. That being said, as a reader when I look at an article such as this one that only displays average FPS when I know for a fact that you can have higher averages and still deliver a poorer experience, I feel like I'm not being told the whole story.

[QUOTE="dirtyferret, post: 1482970, member: 366636"
..and then posted a link to an "article" that declared a winner without actually providing the whole story, the difference being the winner is the brand you root for.

When does Alex ever declare a winner? He stated in his review that it (the 7650K) was a top value, and recommend it for users on a budget, which it is in fact a great option for. The data he posted showed it beat the G3258 in a variety of tests, and he did not exclude data to make one product look better than another.

Allow me to correct you in all the things you fail to understand.

I've been building PCs for longer then you have been alive and I have worked with various brands across the board. That hardly gives me a have a right to state what data is important, a fan boy who claims he is an amateur reviewer for a no name web site with no traffic has just as little right. You see if you don't like it you can speak up and request certain data to be added but to attack the writer is ***** fan boy tactics at best.

Your linked "site" clearly prefers the 7650k as stated by you even if every professional reviewer (people who are paid for talent or service) have posted data otherwise. Readers can make their own judgement based on the facts provided by well known web sites or fan boy sites. You are entitled to your opinion (however flawed) and entitled to your data request. Coming to this site an whining as you have, I've seen more maturity from my infant daughter.

 

[Donny Stanley]

Allow me to correct you in all the things you fail to understand.

I've been building PCs for longer then you have been alive and I have worked with various brands across the board. That hardly gives me a have a right to state what data is important, a fan boy who claims he is an amateur reviewer for a no name web site with no traffic has just as little right. You see if you don't like it you can speak up and request certain data to be added but to attack the writer is ***** fan boy tactics at best.

Your linked "site" clearly prefers the 7650k as stated by you even if every professional reviewer (people who are paid for talent or service) have posted data otherwise. Readers can make their own judgement based on the facts provided by well known web sites or fan boy sites. You are entitled to your opinion (however flawed) and entitled to your data request. Coming to this site an whining as you have, I've seen more maturity from my infant daughter.

You speak of maturity and yet you've had to sensor yourself several times as you decided to insult me, my colleague, as well as the outlet I work for. You call me a "fan boy" because I've criticized the testing methodology of an article which declares one product better than another, however If you've bothered to look at that no named site, you'd notice that myself and other writers have reviewed products from AMD, Intel and nVidia and have always judged them on their own merits. I personally use an Intel system for my daily uses, and an Intel 5960X as my standard test bench. I have absolutely no loyalty to one brand or another. Also, based on your comments, you're clearly quite happy with the G3258 being declared the winner, and seem to be upset about the fact that I am questioning its merit, one might call that being a "fan boy"

You accuse me of attacking the writer, but my original comment merely asked why the data was omitted and stated that the lack of it seemed disingenuous. Which, if you know anything about game performance tests you'd know that your average FPS can be high and still deliver a stutter-y, poor experience.

 

[Steve]

Hi Guys, for those of you who want to see minimum frame rates I have some results for you. Right now I am buried in Batman: Arkham Knight benchmarking so I wasn’t able to provide results for all the games, especially those on my steam account as it is being used for Batman testing.

Anyway here are the results, they are as expected.

Please note for measuring minimum frame rates I took the average of six runs, not the normal three. I have found when measuring minimums you need twice as much data to ensure accuracy.

As we said in the review for a lot of games the performance is much the same when comparing these two budget processors. However the inconsistency of the Athlon X4 860K was a problem in a few games which is why we preferred the Pentium G3258. Games such as Sleeping Dogs, ARMA 3, Thief, Total War Attila and Assassin’s Creed Unity all played very poorly with the 860K.

 

[dirtyferret]

Allow me to correct you in all the things you fail to understand.

I've been building PCs for longer then you have been alive and I have worked with various brands across the board. That hardly gives me a have a right to state what data is important, a fan boy who claims he is an amateur reviewer for a no name web site with no traffic has just as little right. You see if you don't like it you can speak up and request certain data to be added but to attack the writer is ***** fan boy tactics at best.

Your linked "site" clearly prefers the 7650k as stated by you even if every professional reviewer (people who are paid for talent or service) have posted data otherwise. Readers can make their own judgement based on the facts provided by well known web sites or fan boy sites. You are entitled to your opinion (however flawed) and entitled to your data request. Coming to this site an whining as you have, I've seen more maturity from my infant daughter.

You speak of maturity and yet you've had to sensor yourself several times as you decided to insult me, my colleague, as well as the outlet I work for. You call me a "fan boy" because I've criticized the testing methodology of an article which declares one product better than another, however If you've bothered to look at that no named site, you'd notice that myself and other writers have reviewed products from AMD, Intel and nVidia and have always judged them on their own merits. I personally use an Intel system for my daily uses, and an Intel 5960X as my standard test bench. I have absolutely no loyalty to one brand or another. Also, based on your comments, you're clearly quite happy with the G3258 being declared the winner, and seem to be upset about the fact that I am questioning its merit, one might call that being a "fan boy"

You accuse me of attacking the writer, but my original comment merely asked why the data was omitted and stated that the lack of it seemed disingenuous. Which, if you know anything about game performance tests you'd know that your average FPS can be high and still deliver a stutter-y, poor experience.

Allow me to correct you again ( it is becoming redundant) but please pay attention as you may actually learn something

1. several defines as more then "one", there is only one word blocked out in my post and it rhymes with udiot, hardly a sensor able word. So you are wrong there in your use of the word "several".

2. I merely stated the obvious about your site, it has little to no traffic. If you don't like it, then build it up to more traffic otherwise that is a fact.

3. your "fan boy" comments speak for themselves and define you as such

If you would like to continue this discussion perhaps we can move over to your site. I noticed the comments to the article you linked are at zero and you can certainly use the traffic but honestly as far as me educating you about PC tech in future posts, I have neither the patience or desire.

Hi Guys, for those of you who want to see minimum frame rates I have some results for you. Right now I am buried in Batman: Arkham Knight benchmarking so I wasn’t able to provide results for all the games, especially those on my steam account as it is being used for Batman testing.

Anyway here are the results, they are as expected.

Please note for measuring minimum frame rates I took the average of six runs, not the normal three. I have found when measuring minimums you need twice as much data to ensure accuracy.

As we said in the review for a lot of games the performance is much the same when comparing these two budget processors. However the inconsistency of the Athlon X4 860K was a problem in a few games which is why we preferred the Pentium G3258. Games such as Sleeping Dogs, ARMA 3, Thief, Total War Attila and Assassin’s Creed Unity all played very poorly with the 860K.

Several other sites have posted similar results, at minimum frame rates the athlon at best kept up with the pentium and at worst fell behind. It looks like the athlon really fell behind in a few games.

 

[Jad Chaar]

Thanks for the benchmark, Steve. I was recently thinking of putting the 860K in my friends rig but I may put the Pentium in it now.

 

[Guest]

Really nice review and comparison. The day has come where Intel's Dual Core beats AMD's Quad Core. Can't believe my eyes, cause here I am using the old Phenom II x4 955 BE, guess it's time to switch sides and never look back at AMD's processors. This is really shameful and pathetic for AMD.

 

[Guest]

Minimum, max and average FPS does not tell the whole story.

Frame time variance/frame latency us where the valuable data is at. That said the G3258 seems to be behind even the 750k in frame latency in several benchmarks until they are overclocked a good amount. That said it's almost 50/50 in people's stances on if the G3258 is a stuttery mess in some games.

Current benchmarks just shows how behind AMD is right now and game/software development in general when games are not coded to efficiently use multiple cores while allowing a pure dual core to beat a quad core.

Hoping AMDs new architecture can at least protect it's CPUs from Intels with lower cores. So don't allow a dual core Intel (w or w/o HT) to beat a Zen Quad-Core, etc.

 

[Untuk]

Ok so from this benchmark test, we could say that Intel got better performance, but all the test that being done, was taken from the gaming perspective. My question is, because I tend to use my PC for less gaming, and more productivity things, multi tasking like, browsing, typing in words, programming (coding), image editing with photoshop, and even audio and video editing, which processor will perform better, or get things done faster?
Is it athlon because it have more cores, or is it still the intel with beter single core performance? Please enlighten me.

 

[Steve]

Ok so from this benchmark test, we could say that Intel got better performance, but all the test that being done, was taken from the gaming perspective. My question is, because I tend to use my PC for less gaming, and more productivity things, multi tasking like, browsing, typing in words, programming (coding), image editing with photoshop, and even audio and video editing, which processor will perform better, or get things done faster?
Is it athlon because it have more cores, or is it still the intel with beter single core performance? Please enlighten me.

This might help...
https://www.techspot.com/review/849-intel-pentium-anniversary-edition-overclock/page10.html

Think of the A10-7850K as the Athlon X4 860K but with an iGPU. So in other words compare the A10-7850K to the G3258.

 

[EEatGDL]

Ok so from this benchmark test, we could say that Intel got better performance, but all the test that being done, was taken from the gaming perspective. My question is, because I tend to use my PC for less gaming, and more productivity things, multi tasking like, browsing, typing in words, programming (coding), image editing with photoshop, and even audio and video editing, which processor will perform better, or get things done faster?
Is it athlon because it have more cores, or is it still the intel with beter single core performance? Please enlighten me.

In theory, multi-tasking and compilations must benefit from the Athlon; encoding shall benefit from the Pentium, since both have the same amount of FPUs and the Pentium's are more efficient. Where the Athlon wins is in tasks that benefit from extra cores for integer operations.

 

[Allan M Systems]

Hello Techspot,

I was just reading though this article, and tests. Thank you for taking the time to perform these tests and share your result's.

I have a few questions and thoughts.

Were there any background tasks running? Like a anti-gunkware suite, or open web browser tabs? Or teamspeak? Steam?

In the articles conclusion, it is mentioned that in most games the 860K achieves 60-90% usage as reported by a hardware monitor. What was the usage on the Pentium?

Are there any provisions in your testing to capture the effects of multi-player gaming conditions, rather than the usual single player test sequences?

The test results of Crysis 3, Bioshock Infinite, Metro Redux, Tomb Raider, Sleeping Dogs, BF4, BFH, Watch Dogs, DA:I, and Witcher 3, all appear to be largely muddled under the vale of significant GPU bottlenecks. Very little CPU related performance scaling is popping up in any of these tests. This tells me that tests were likely performed using single player tested sequences (where applicable) or, with visual quality settings set too high to reveal the difference in compute performance.

---------

In almost any raw CPU compute test, the 860K proves to have more execution resources, and a higher maximum execution throughput than the G3258. For CPU gaming benchmarks, the arrangement of execution resources in the G3258 provides higher execution throughput to the workload as presented. I don't think there's any disputing that the G3258 produces better looking results in single player test sequences on clean system installs, and in fact, if you were to lift the GPU bounding vale imposed on most of your test results, the G3258's lead would open up even further. I'd really like to see an attempt made to use multi-player conditions and/or lower visual quality settings where applicable in those titles mention in #4 above to shift the bottleneck back the CPU, to get a more realistic picture of the real difference in performance between these CPU's in gaming.

Thank You,
Eric