AMD FX 8350 Winning against i5/i7?

[SPBHM]

I would love to see Anandtech perform a similar test because the methodology and documentation performed by this guy in the video are not impressive at all... The results do seem to be all over the place and I heard no mention on the quality of the output stream. With such low FPS in some cases though, it looks like he gave Xsplit priority as far as CPU cycles are concerned though but that is just a guess.

why this?

Crysis Warhead xsplit
1080 - 26.44
1600x900 - 39.28
720 - 48.28
Crysis Warhead
1080 - 35.64
1440 - 26.772

to me it looks like they are trying to mix results with and without xsplit being used!?

and some of the results make no sense at all, with or without using xsplit.

 

[guskline]

I've noticed something a bit strange about this forum and in particular postings where the FX 8350 is compared to the Intel counterparts of I5/I7 (2500k/2700k/3570k/3770k). When someone runs a test between these chips and the 8350 scores slightly higher on a test, posters seem to immediately come out of the wood work attacking the credibility of the test, the tester etc, yet when the results favor the Intel chips no one questions the tester, the test etc.

Why? Isn't it possible that under certain parameters the 8350 actually does produce ma higher result?

I watched this video by the tester who ran xsplit and ran a few games and said he was amazed that the 8350 actually beat the 3570k, in particular by a few fps so he ran the test again and again. I think his mannerisms are somewhat quirky, but he appears to be a legit tester. Why the attacks?

BTW, I have said again and again, for gamers, all things being equal the 3570k might be a better buy than the 8350. However, the overall performance of the chips is close AND in some software scenarios the 8350 actually pulls ahead.

 

[guskline]

why this?

Crysis Warhead xsplit
1080 - 26.44
1600x900 - 39.28
720 - 48.28
Crysis Warhead
1080 - 35.64
1440 - 26.772

to me it looks like they are trying to mix results with and without xsplit being used!?

and some of the results make no sense at all, with or without using xsplit.

I agree that the video appears somewhat disjointed, but the results above appear to make sense. While streaming with xsplit the fps climb if you use a lower resolution, while if you just run the game without streaming you get the results posted below that.

 

[Idontcare]

I've noticed something a bit strange about this forum and in particular postings where the FX 8350 is compared to the Intel counterparts of I5/I7 (2500k/2700k/3570k/3770k). When someone runs a test between these chips and the 8350 scores slightly higher on a test, posters seem to immediately come out of the wood work attacking the credibility of the test, the tester etc, yet when the results favor the Intel chips no one questions the tester, the test etc.

Why? Isn't it possible that under certain parameters the 8350 actually does produce ma higher result?

Extraordinary claims require extraordinary proof. It is extraordinary to claim the FX8350 is outperforming a processor that costs $100 more.

As such, it is not unusual to have people be suspicious of the results until extraordinary proof is provided to justify the extraordinary claim.

The method in which this suspicion is communicated though is just your standard human trait we call "knee-jerk reaction" and nothing more. Sure it could be communicated a little more diplomatically on occasion, but the general message is still the same.

 

[Puppies04]

I was on the market for a mid range CPU. I was originally gonna get a i5, but then I saw this video. This is very interesting results

Youtube Link

None Video Link: http://teksyndicate.com/videos/amd-...s-3820-gaming-and-xsplit-streaming-benchmarks

Tl;dr FX 8350> i5 9/10 times



Could this be a gpu bottleneck, should I stick with the i5? I am actually considering getting a HD 7870.

What the hell is that dufus in the link talking about at 2:10? "Well lets talk about the different architecture, the intel chip has a large pool of shared L3 cache and so does the FX but the FX has seperate L2 cache for each of its 4 cores.... erm hello so do intels chips. Turned it off at that point.

 

[frozentundra123456]

I've noticed something a bit strange about this forum and in particular postings where the FX 8350 is compared to the Intel counterparts of I5/I7 (2500k/2700k/3570k/3770k). When someone runs a test between these chips and the 8350 scores slightly higher on a test, posters seem to immediately come out of the wood work attacking the credibility of the test, the tester etc, yet when the results favor the Intel chips no one questions the tester, the test etc.

Why? Isn't it possible that under certain parameters the 8350 actually does produce ma higher result?

I watched this video by the tester who ran xsplit and ran a few games and said he was amazed that the 8350 actually beat the 3570k, in particular by a few fps so he ran the test again and again. I think his mannerisms are somewhat quirky, but he appears to be a legit tester. Why the attacks?

BTW, I have said again and again, for gamers, all things being equal the 3570k might be a better buy than the 8350. However, the overall performance of the chips is close AND in some software scenarios the 8350 actually pulls ahead.

In anands bench, out of 30plus benchmarks the 8350 wins 4 and 1 is a tie vs the 3770. Additionally, most of the wins are by a relatively small amount. It seems reasonable to me to question benchmarks which show the 8350 ahead by as much as this video does vs the 3770. Your post only mentions the 3570. Against that chip the 8350 is ahead in a fair amount of heavily mt benchmarks, but only in a very few is it ahead of the 3770.

 

[bononos]

I've noticed something a bit strange about this forum ...... BTW, I have said again and again, for gamers, all things being equal the 3570k might be a better buy than the 8350. However, the overall performance of the chips is close AND in some software scenarios the 8350 actually pulls ahead.

The 'strange' thing happens because there is a stubborn few that refuses to believe that all things being equal 3570k is clearly superior for gamers, not just 'might be better'.

 

[SlowSpyder]

I've noticed something a bit strange about this forum and in particular postings where the FX 8350 is compared to the Intel counterparts of I5/I7 (2500k/2700k/3570k/3770k). When someone runs a test between these chips and the 8350 scores slightly higher on a test, posters seem to immediately come out of the wood work attacking the credibility of the test, the tester etc, yet when the results favor the Intel chips no one questions the tester, the test etc.

Why? Isn't it possible that under certain parameters the 8350 actually does produce ma higher result?

I watched this video by the tester who ran xsplit and ran a few games and said he was amazed that the 8350 actually beat the 3570k, in particular by a few fps so he ran the test again and again. I think his mannerisms are somewhat quirky, but he appears to be a legit tester. Why the attacks?

BTW, I have said again and again, for gamers, all things being equal the 3570k might be a better buy than the 8350. However, the overall performance of the chips is close AND in some software scenarios the 8350 actually pulls ahead.

I think some posters are so used to the idea of Intel being the dominant CPU maker, that they cannot get their heads wrapped around the idea that sometimes, in some scenarios, the FX may be a faster part than the Intel chips. Also, I think a lot of the time a poster automatically thinks of overclocked CPU vs. overclocked CPU. Intel really shines when both are overclocked, but a lot of people do not overclock.

I am with you, if you look at the CPU's put out by both companies, Intel overall makes the better chips. Not much to argue there. But, sometimes AMD can be a better choice, the FX is good enough to at least consider depending on your situation. That wasn't the case with Bulldozer. I think many have automatically associated anythng FX or related to Bulldozer as inferior. If AMD could just do somethig about power consumption, I think they would have a really solid part.

 

[guskline]

Is there something in the coding of the xsplit software that would favor the 8350 over the 3570k/3770k?

 

[coffeejunkee]

I think some posters are so used to the idea of Intel being the dominant CPU maker, that they cannot get their heads wrapped around the idea that sometimes, in some scenarios, the FX may be a faster part than the Intel chips. Also, I think a lot of the time a poster automatically thinks of overclocked CPU vs. overclocked CPU. Intel really shines when both are overclocked, but a lot of people do not overclock.

Nah it's not that, it's just that most of the results are really strange. I'm not surprised the 8350 does better if you game and run a heavily multithreaded app at the same time. But the results without Xsplit are just weird, Crysis Warhead for example:

Crysis Warhead, 1080p
8350:
35.64
3570K:
26.84
3770K:
38.44
3820:
26.84

Why is the 3570K doing so bad? Crysis Warhead doesn't need lots of threads and is very gpu dependent. HT and 100MHz faster clockspeed shouldn't give the 3770K 12fps advantage. And if it does, why is the 3820 doing so bad, it's practically the same cpu as a 3770K.

And then at 1440p you would expect a big fat gpu bottleneck with a single HD7870 but there is still a lot of difference:

Crysis Warhead, 1440p
8350:
26.772
3570K:
18.720
3770K:
23.880
3820:
not tested

Doesn't make sense at all, especially if you look at this http://www.anandtech.com/show/5771/the-intel-ivy-bridge-core-i7-3770k-review/7 where the AMD chips are significantly slower (Ok, low resolution to test cpu dependency and 8150 instead of 8350, but still)

Also, power draw is conveniently ignored and claiming all 8350's can do 5GHz easy is just bollocks.

 

[Greenlepricon]

What I'm wondering is why the resolution changes cause inconsistent fps of each cpu if it's a gpu bottleneck? Granted I'm not too computer savvy by this forum's standards, but shouldn't the fps's all be essentially the same, with possibly small variances? I know that resolution probably does impact the cpu in some ways, but I was under the impression that the cpu will only run a game so well, which can be tested at lower settings, and beyond that the gpu becomes the problem.

Don't get me wrong, I would be really happy if AMD really did pull this lead because it would mean they're not out of the game. I'm just curious why the results would show up like this.

Doesn't make sense at all, especially if you look at this http://www.anandtech.com/show/5771/the-intel-ivy-bridge-core-i7-3770k-review/7 where the AMD chips are significantly slower (Ok, low resolution to test cpu dependancy and 8150 instead of 8350, but still)

Also, power draw is conveniently ignored and claiming all 8350's can do 5GHz easy is just bollocks.

The main problem is that the 8350 is generally MUCH better than the 8150. Should've linked to Vishera review. Also I've heard quite a few of these chips hitting 5GHz, but you're right that it's not always easy and the power draw can get a little outrageous.

 

[coffeejunkee]

The main problem is that the 8350 is generally MUCH better than the 8150. Should've linked to Vishera review. Also I've heard quite a few of these chips hitting 5GHz, but you're right that it's not always easy and the power draw can get a little outrageous.

Well, it has 400MHz higher base clockspeed so it should be. But unfortunately Crysis Warhead isn't tested in that review.

 

[Trm8r]

I've noticed something a bit strange about this forum and in particular postings where the FX 8350 is compared to the Intel counterparts of I5/I7 (2500k/2700k/3570k/3770k). When someone runs a test between these chips and the 8350 scores slightly higher on a test, posters seem to immediately come out of the wood work attacking the credibility of the test, the tester etc, yet when the results favor the Intel chips no one questions the tester, the test etc.

I think its the herd mentality, and we're all taller when we're standing on someone else. We can all go along, with hating AMD, or we can go to h-ll. Many replies to posts, that don't conform with "AMD sucks", start with an accusation and then a personal attack.
I like to think that all of the haters will be sorry if they manage to make Intell and Nvidea monopolies. But who am I kidding, they'll just find a new target for their contempt.

 

[coffeejunkee]

Yes, all hardware sites are wrong. But luckily for us someone made a Youtube video to show us what's really going on.

 

[bunnyfubbles]

Is there something in the coding of the xsplit software that would favor the 8350 over the 3570k/3770k?

Xsplit uses x264 which is heavily multhithreaded and Bulldozer has always "excelled" in x264 (and by excel it just means its not completely terrible and will be, at best, almost as fast as an i7 clock for clock in the 2nd pass)

ultimately I'd have a really hard time believing the 8350 can be better than an i7 3770K for streaming unless the potential user is completely unwilling to overclock, and all it should take to guarantee the 3770K as the superior part (inferring from x264 benchmarks) is to have the automated overclocking most motherbaords have these days get the 3770K up to 4-4.2GHz which nullifies the 8350's clockrate advantage that gives it a slight edge.

Although it might not even take that, as I found it suspicious that they opted not to provide 3770K results in their streaming tests.

 

[Abwx]

Crysis Warhead for example:

Yes , good exemple....

 

[Idontcare]

I think its the herd mentality, and we're all taller when we're standing on someone else. We can all go along, with hating AMD, or we can go to h-ll. Many replies to posts, that don't conform with "AMD sucks", start with an accusation and then a personal attack.
I like to think that all of the haters will be sorry if they manage to make Intell and Nvidea monopolies. But who am I kidding, they'll just find a new target for their contempt.

You probably don't see the irony in your post, do you?

Your post contains more outrage (against whatever) than the posts made by the so-called haters, the same posters for which you so clearly target your contempt.

 

[coffeejunkee]

Yes , good exemple....

Sorry, what's your point exactly?

 

[dragantoe]

if you are gaming go with the 8350, I watched that teksyndicate video earlier, the 8350 will basically blow the i5 away in 90% of games and can be overclocked to an extent the i5 will never see (5.5 ghz with water cooling easy), all with lower temps

 

[ShintaiDK]

if you are gaming go with the 8350, I watched that teksyndicate video earlier, the 8350 will basically blow the i5 away in 90% of games and can be overclocked to an extent the i5 will never see (5.5 ghz with water cooling easy), all with lower temps

The word of the day in that video, xsplit.

So no, get the i5 for gaming.

 

[Sable]

if you are gaming go with the 8350, I watched that teksyndicate video earlier, the 8350 will basically blow the i5 away in 90% of games and can be overclocked to an extent the i5 will never see (5.5 ghz with water cooling easy), all with lower temps

AHAHAHAHAHAHAHAHAHAHAHAHAHHAA

no.

edit:

Wait, you're being sarcastic aren't you.

 

[tweakboy]

I was on the market for a mid range CPU. I was originally gonna get a i5, but then I saw this video. This is very interesting results

Youtube Link

None Video Link: http://teksyndicate.com/videos/amd-...s-3820-gaming-and-xsplit-streaming-benchmarks

Tl;dr FX 8350> i5 9/10 times



Could this be a gpu bottleneck, should I stick with the i5? I am actually considering getting a HD 7870.

Ok hold up one topic at a time LOL. AMD has 8 slower physical cores and Intel has 8 logical cores. AMD is slower in single threaded way behind.

I think a i5 with a 7850 is a mistake. Because the 7850 is pretty darn strong.. Your only dishin the GPU your 2 cores or 4 cores..................... gl

 

[Jovec]

Xsplit uses x264 which is heavily multhithreaded and Bulldozer has always "excelled" in x264 (and by excel it just means its not completely terrible and will be, at best, almost as fast as an i7 clock for clock in the 2nd pass)

It has to do with HT vs CMT, optimized algorithm code, and optimized threaded implementation. Video encoding is essentially a best case scenario for AMD's FX line.

 

[NTMBK]

It has to do with HT vs CMT, optimized algorithm code, and optimized threaded implementation. Video encoding is essentially a best case scenario for AMD's FX line.

It's kind of funny looking at this from a historical perspective. I recently dug through a load of old Anandtech Pentium 4 era articles (don't ask), and it was always Intel which excelled on video encoding and failed on gaming compared to AMD's Athlon 64. How times have changed.

 

[dragantoe]

The word of the day in that video, xsplit.

So no, get the i5 for gaming.

he did gaming and xsplit benchmarks