AMD FX-8350 Review: Does Piledriver Fix Bulldozer's Flaws?

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.
G

Guest

Guest
[citation][nom]cobra5000[/nom]jbx007, there is ALWAYS a mark up, when the hot new product comes out. It dies down pretty quickly, as long as the supply meets the demand.[/citation]

i totally agree, but my post took into account that intel will be responding within days, like they did last time, to announce a price cut to offset any gains that amd my have had if they had the under $200 price.

at least that is what happened to me last time, i went with amd the day it came out, actually the very next morning, put together a new system, paid the markup and dealt with the problems with the not yet updated motherboards bioses and then within that very week intel dropped the prices bigtime, if i would have waited i would have gotten a much better deal with intel and tested chips and motherboards. anyways just my thoughts

i guess like i said, ill stay on the fence waiting to see what intel does and what problems amd new chip / motherboards have.

to be honest to this day those early motherboards still give me trouble... :(

also, dont get me wrong i've had similar problems with motherboards when intel has released new chips and chipsets. and both have cost me nothing but pain trying to work around beta bioses ;)
 

mayankleoboy1

Distinguished
Aug 11, 2010
2,497
0
19,810
Would FX-8350 be my first choice in a new build, though? Probably not. Although I’m impressed by the work AMD’s architects have done in the last year, performance remains too workload-dependent.

This sentiment probably wont ever change in favour of AMD. They are too far behind performance and power. But unless this changes to "probably yes" , i dont see AMD grabbing majority of the market share.
Chris :Now that PD is out, and AMD has given details of Steamroller, what is your personal estimate of the performance improvement of Steamroller ?

 

Hazle

Distinguished
*slowly clap*

well done, AMD. a low bar over BD, granted, but a congratulation is definitely in order. based on the few FX-4300/6300 reviews elsewhere so far, i can finally stop recommending Phenom II's for gaming purposes. i may even consider getting a 6300.

VIVA LA COMPETICION!

.... or is it competencia?
 
G

Guest

Guest
AMD moved forward here for sure.

My 3570K clocks to 4.8Ghz on air and still uses less power!

For gaming, it is still faster, but AMD have moved forward!
 

mayankleoboy1

Distinguished
Aug 11, 2010
2,497
0
19,810
The gaming benchmark methodology is somewhat irrelevant. The "average FPS" is good for lots of the CPU's. 45 Vs 60 is not much significant.
But the frame latencies (ala techreport) are more relevant and informative. Of course, they are harder to grasp.
 

tomfreak

Distinguished
May 18, 2011
1,334
0
19,280
Looking at how poor AMD IPC(or per core performance) is. I dont think I will be buying AMD for long long time. I personally like high single threaded performance CPU.
 

proffet

Honorable
Aug 30, 2012
489
0
10,810
[citation][nom]looniam[/nom]sorry just not overly impressed.5-12% performance increase 12% less power - sound familiar?the only difference this time was less hype before the release. (lesson well learned AMD!)[/citation]
I'm with you on this 'looniam'...

[citation][nom]blazorthon[/nom]Honestly, I'm disappointed in Vishera. Comparing it to Trinity, it seems that the L3 cache doesn't actually make a difference in performance for these chips. Maybe its L3 cache's latency is simply too high for it to do much of anything other than suck power. Some CPU/NB frequency overclocking tests might be able to confirm this and if so, solve the problem and let Vishera really pull ahead of Zambezi and Trinity.[/citation]
maybe some hope for you yet bro, I actually like this reply from you.. :)
 

proffet

Honorable
Aug 30, 2012
489
0
10,810





+1 - been saying that since Bulldozer...


and still should be better at all of the above...


and all this sums it up for me.
see you around Steamroller or whatever the next name is.. :/
in-store only deals @ Fry's and/or Microcenter (if you live near one) and price match a 3570K for $195 with tax out the door.
 
G

Guest

Guest
I know Newegg is your official pricing partner, but I was able to pick up a 3570K at a local micro center for $190 this summer. The FX8350 should be priced around $180 IMP. Yup just went to their website and they're still showing the 3570K for $189. So I think the Newegg price is suspicious.
 
considering all i really do is game and browse the web, im glad to stick with my 3570k, no regrets here. Really, what percentage of people do things that take advantage of 8 threads that would make this cpu worth a look? maybe 10%? People are happy that there is a decent improvement here, but imagine what they could have done if they had stuck with tweaking the phenom II core? and stuck 8 of those cores on a cpu? They would have a much better product today and would be much more competitive i think. Instead they went the pentium 4 route...... and we all know where that leads.......
 

adgjlsfhk

Honorable
Feb 21, 2012
518
0
11,010
Great review. AMD's biggest problem by far is their power consumption. By now, they have pretty good performance even com[pared to ivy, but 125 watts will not compete with 77 when it comes down to it.
 
G

Guest

Guest
Can we please stop with the "8 cores can't beat Intel's 4 cores" nonsense when most of the benchmarks aren't using all 8 cores? It's this endless strawman argument that has nothing to do with the reality of the situation.

The benchmarks where AMD are winning tend to use all 8, the benchmarks that are close are
 

silverblue

Distinguished
Jul 22, 2009
1,199
4
19,285
Yes, but even so, if even one core per module is being hit, it cannot go idle on the parts that aren't being properly used, meaning unnecessary power consumption. At least, as far as I remember. Besides, if all eight cores are in use and not to their full potential, it's still using all eight cores and burning power unnecessarily. Tiny integer cores are still integer cores, after all.
 

mikenygmail

Distinguished
Aug 29, 2009
362
0
18,780
4 ghz 8350 CPU stock, and 5 ghz when overclocked? Very nice.
More than enough performance for any single threaded applications and games,
and fantastic 8 core performance for multithreaded applications and games.
 
Not bad for the price but 10% boost just from clocks alone isn't good, what people are wanting and for some really need is performance per clock. Guess that I will be sticking with my p2 x4 820 with L3 unlock for another year :s
 

proffet

Honorable
Aug 30, 2012
489
0
10,810

your lucky to get the L3 to unlock, mine didn't when I had the P2 X4 840... I was disappointed to say the least.. :/
 

mikenygmail

Distinguished
Aug 29, 2009
362
0
18,780
What many reviewers and fanboys tend to miss over and over again is that AMD delivers the best performance-per-dollar and that ANY current model desktop CPU will run ANY software just fine. Unless you have some enterprise level software that brings a modern CPU to it's knees, ANY of the currently available desktop CPUs will run Windows or Linux based software just fine. In fact, Linux apps do even better in many cases than Windows bloatware.

I have no idea if AMD will ever offer a discrete CPU to equal Intel's top of the line, over-priced models nor do I care. I buy what delivers the best performance for the price. I have yet to purchase any AMD desktop CPU that would not run ALL software as well as an Intel CPU, without any isses what so ever.

If all you do is benchmark all day long and you have money to burn, blow it on an Intel CPU, unless of course you are opposed to evil, chronic, law violating corporations looking to eliminate consumer choice. You could always vote your conscience, if you have one.

I am always amazed that people actually believe, albeit falsely, that AMD processors are some how "inadequate." Even with tainted benches, AMD processors deliver all the performance and good value that most consumers desire. It's tough however getting people to look at the data objectively. All most people think of is that "more" is better, when in fact that's the sucker play when you look at performance vs. cost and actual needs.

Considering that Intel got a whopping ~5% performance gain from a 32nm to 22nm node drop and Tri-gate tansistors with Ivy Bridge, (along with over-heating and poor overclockability...) AMD did quite well to deliver a ~10-15% improvement with Vishera. With AMD's pricing Vishera should sell well, since it delivers excellent performance at a low price.
 
I dont understand why there is no SC2 test. Its by far the most impacted game by CPU performance, and is also a modern game.

I think getting 3 Graphic dependant games (and 1 of them beeing a port from console version) is not a good example of carefull testing.
And while i do understand that those are the most popular games at the moment, it simply dosent give enought information when it comes to the gaming department.
 

army_ant7

Distinguished
May 31, 2009
629
0
18,980
Nevertheless, FX-8350’s loss to FX-8150 is a result of that lower data rate. Had we used DDR3-1866 modules, the FX-8350 would have matched the Ivy Bridge-based chips at DDR3-1600 closer to 20 GB/s.
Is this a margin of error thing or does Vishera have a slower NB frequency (if that matters at all) or something? :)

I do remember hearing as well that Windows 8 runs BD better than a patched Windows 7. Was this false, or was there a newer, improved patch for Windows 7 that came out or did they just figure out how to make Vishera work better that Bulldozer with the old (but last known (to me)) patch? Maybe a CPU driver (automatically through Windows Update) or firmware adjustments (Hm... CPU's do have firmware right?)? :)

I wouldn't call it an overclock if those improvements to Vishera allowed it to hit higher clocks more efficiently than Zambezi. I'd call the raised stock clock as part of the architecture's features.
 

Mathos

Distinguished
Jun 17, 2007
584
0
18,980
It has occurred to me that even after 6 years, they still haven't figured out that the biggest bottleneck with AMD's processors is the Memory controller, Northbridge and L3 cache speed all being the same as the hypertransport speed. Through my own experimentation with Previous gen Phenom 9600be and 9850be processors I could prove that upping those frequencies to match the core speed improved performance, even at stock core speed. In general it lowered L3 latency considerably over stock. Obviously this allowed the cores to be fed data more quickly and with less wait. Almost wish I could still find my old 9600\9850be screenshots to display this. It baffles me, that if current hyper transport specs allow for 3.2Ghz 6.4GT/s, why are they still continuing to run their nb/imc/L3/HT speed at 2-2.1ghz or 4-4.2GT/s on the HT side. For gods sake, they need to at least run it up to 2.5ghz 5GT/s to see if it closes any gap with the 3rd gen i series chips.

Even current 3rd gen i5's and i7's run their DMI interface at 5GT/s. But, I'm gonna be the L3 cache on those chips isn't tied to the speed of the memory interface though. It appears for intel the L3 is tied very closely to the actual core speed.

I'm gonna say that's the most likely reason AMD lately has historically bottlenecked video cards when it comes to many games lately. At least since the first k10 chips.

Now on to the 8350, bout time, at least now it looks like I have a compelling reason to upgrade from my current 1090t. Especially considering video encoding with h.264 is one of the things I do.
 
Status
Not open for further replies.