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

Page 9 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.
Wow... Many changes indeed. It makes me excited to play again. Though maybe next time. I've also heard of something to do with a new combat system (I think). Look into it if you're interested. :)

I was aware of those two quest changes. I forgot though what the Romeo and Juliet quest was replaced with though. Well, anyway, we're deviating from the topic of this thread. Hehe... I appreciate all that info. 😀 Feel free to PM me your user name on RS if ever you start playing again (though I don't guarantee that I will when you do). :)
 
nice job by amd, but they need to reduce power consumption. like instead of using a tank circuit, they should use a KERS like technology...a tank circuit is good because it stores electrical energy. electricity passes from the capacitor to the inductor, where it creates a magnetic field. When the capacitor’s charge reaches zero, the current flow reverses — power shifts from the inductor to the capacitor, and the magnetic field dissipates but KERS is much better, because it also stores electricity, the difference is instead of electricity passes from the capacitor to inductor/ vice versa, where it creates a magnetic field, then the current flow reverse when the capacitors charge reaches zero... (KERS)Once the energy has been harnessed, it is stored in a battery and released when/where it is needed.
 
The only thing i see is its right in there like a dirty shirt for all the battlefield 3 FPS compared to the intels. Good job. Its very tempting to get and hell you get 8 cores... Eventually you will need them once games etc catch up to it.

I can see this being a very good gaming machine.
 
guys huge power usage, core2duo single thread performance

(dont believe me google anandtech bench)

nooooo

do yourself a favor $120 IVB core i3 native USB3 and PCIe3.0 or PII X4 if you must have AMD

amd burned all their $$ on ati and did no r&d so they are trying to sell us amp'd up server parts
 


These FX processors have better multi threaded performance than i5s and almost as good as i7s. This is really important because single threaded applications are quickly becoming a thing of the past. So this processor is "future proof" but still lacks in single core apps that are still present in games or outdated programs. This processor also overclocks like a champ and you can address poor single threaded performance that way.

The only real draw back is that its power consumption is still very high when compared to intel ships.
 
[citation][nom]calguyhunk[/nom]So when do we get the benchmark comparos with the Phenom IIx4 9xx range?[/citation]

the 3/4 year old PII X4 9xx still strong ^
 
[citation][nom]slippyrocks[/nom]guys huge power usage, core2duo single thread performance(dont believe me google anandtech bench)nooooodo yourself a favor $120 IVB core i3 native USB3 and PCIe3.0 or PII X4 if you must have AMDamd burned all their $$ on ati and did no r&d so they are trying to sell us amp'd up server parts[/citation]

FYI, the Ivy i3s don't have pcie 3.0 support.

http://ark.intel.com/products/65693/Intel-Core-i3-3220-Processor-3M-Cache-3_30-GHz
 
I only build a new system if & when I can get 2x the performance of my current system. Looks like that is still a long way off. On the plus side, by the time this happens AMD may once again be the King of Performance :) OR they may have left off making desktop components all-together.
 
if they keep on improving their processors they could be competitive again... hope they dethrone INTEL processors... hahahahaha :)
 
^ They are already competitive with these new line of processors. They are cheap, have superior multi threaded performance compared to i5s and very close to i7s, they overclock really well, you get 6 to 8 cores for excellent multi tasking performance, such as multiple displays or programs, and if that wasn't all they are dirty cheap.

Intel motherboards and i7s are much more expensive so for any person on a budget the FX processors are an excellent upgrade, not to mention that the Steamroller will also be amd3+ so you get to keep your mobo.
 
[citation][nom]oxford373[/nom]thanks chris nice review, but can you consider to add benchmarks for metro 2033 since it is the only highly multithreaded game available in the next 6300/6400/8320 review http://www.tomshardware.com/review [...] 120-8.html because this is the only game that allowed fx 8120 to outperform fx 4100 and fx 6100,and PII x6 1090t to outperform PII X4 980.anyway i found benchmarks for this game for fx-8350. http://www.overclockersclub.com/re [...] 8350/8.htm[/citation]

That looks to me like it's more of a wash between those CPUs that is caused by a GPU bottle-neck than actually having them scale across many threads effectively. That the i5-2500K, i7-3820, and the i7-2600K had identical performance and the overclocks hardly change anything on any of the CPUs shows this quite convincingly IMO. Heck, that the i7-3960X is around the end of the pack is clear proof of it.
 
Though we already know the reason why multi-player in Battlefield 3 isn't used as a benchmark (yet) here, it is also highly threaded I hear.
 
Finally, an improvement to AMD's non-APU lineup. It was many years in the making but it looks like they're finally offering a compelling CPU again. Aside from power, I guess my biggest concern for AMD is the pricing. Years ago when they conceived the latest FX series, they were expecting to get good prices for their chips (better margins in other words). If there is one thing Intel's superiority has done to the PD-based FX line, it is to require them to be affordable. This no doubt hurts AMD's bottom-line and I doubt they will be selling enough volume until maybe Steamroller to make up for the loss of projected revenue due to lower-than-expected initial pricing (remember BD's unreasonable INITIAL price points? They obviously learned from that mistake).

I'm just hoping AMD will be able to stay afloat considering their lack of high-end performance and lower revenues from stiff competition. A 4-core Intel +HT easily beats an "8-core" PD chip using half the power - this is definitely a problem for AMD going forward, but with their competitive pricing in the low and mid-range, the newest FX chips are indeed encouraging!

Oh, and AMD seriously needs a new socket - the fact that they are trying to get board makers to issue BIOS updates for 4, 5, 6 year-old boards with new procs is completely ridiculous - this is holding them back and creating major headaches for people buying AMD boards. The board makers hate it. As mentioned in the article, some mobo makers have stopped issuing updates on select boards entirely since those boards are well-past their standard support cycle - this will be a major problem going forward. It's one of those, "Cross your fingers and maybe your motherboard will have an update so you can use PD or SR." It was the same with BD, a number of people thought their beat-up old boards would be compatible. Sure it was ultimately their fault for not checking with their mobo maker, but I felt like a doctor telling a patient terrible news - "Sorry your beat-up old board won't run BD despite what you heard from AMD" "But I already bought the chip!"...
 
I think AMD did some good stuff here. In it's threaded performance it did very good and it may end up getting them some server business that they really need. They did nice, it would have been nicer to see them punch up the cache latency as well but their performance (even if it's a bit hotter) gives them something to smile about. The numbers are respectable this time and with their price they set them up in a good range to be considered a good deal. With the APU's just maybe AMD might see their profits go up and this they really really need if they plan on being around in the future.
I hope they get the needed support, the numbers show it to be a pretty good cpu and with the software going more and more towards threading it is a cpu that considers the future built in. It was just the past that really hit them with the single threaded software in the first place.
They should have paid a bit more attention to that but at least this new cpu has shown improvements all the way across the board so they did do a better job and it does show it this time.
 
[citation][nom]army_ant7[/nom]Though we already know the reason why multi-player in Battlefield 3 isn't used as a benchmark (yet) here, it is also highly threaded I hear.[/citation]

maybe all games are multithreaded in multiplier,but who cares if we cant find any benchmark for any game in multiplier mode(or we cant prove that).

@blazorthon, i think metro 2033 prove like many other games that games benefit from AMD modules like real cores when almost all games don't benefit from hyperthreading technology for example,despite metro 2033 is highly multithreaded game,i3-2100 3.1ghz was just 2 frames faster than Pentium G860 3.0ghz,when fx 8120 outperforms fx 6100 and fx 4100.
http://www.tomshardware.com/reviews/gaming-fx-pentium-apu-benchmark,3120-8.html
 
[citation][nom]oxford373[/nom]maybe all games are multithreaded in multiplier,but who cares if we cant find any benchmark for any game in multiplier mode(or we cant prove that).@blazorthon, i think metro 2033 prove like many other games that games benefit from AMD modules like real cores when almost all games don't benefit from hyperthreading technology for example,despite metro 2033 is highly multithreaded game,i3-2100 3.1ghz was just 2 frames faster than Pentium G860 3.0ghz,when fx 8120 outperforms fx 6100 and fx 4100.http://www.tomshardware.com/review [...] 120-8.html[/citation]

It's easy to prove that BF3 MP is extremely CPU-intensive and that it benefits greatly from highly threaded CPUs. Benchmarking to prove that is easy. What is difficult is making benchmarks that are consistent enough to have multiple runs compared without extreme outliars. Basically, the accuracy of comparisons is extremely poor. You can make generalizations accurately, such as that BF3 loves threads, but you can't accurately get numbers for exactly how much of a difference they make nor comparing CPUs of different micro-architectures. Given the inaccuracy, it also takes a lot of runs to try to average out the outliars and it's a hassle to spend a lot of time benchmarking a single game when in that same amount of time you could have benchmarked several other games.

As for the HTT versus real core difference theory in Metro 2033, maybe you're right, but I'll again point out how the i7-3960X didn't agree with it. Maybe it's simply an exception because otherwise, this generalization seems to be quite accurate, but IDK. I haven't done a thorough Metro 2033 CPU comparison and honestly, I haven't really looked into it. It's such a GPU limited game that many different CPUs perform similarly even if they perform vastly different in most other games.

Also, have a look at this:
trinity-99th-discrete.png


Hyper-Threading might not have a big impact on FPS, but in frame latency aka closer to what you actually see, it seems to have a huge impact for Intel's dual-core CPUs. Maybe this isn't true in Metro 2033 (that game isn't part of the review that I got this from), but it seems to generally be true even for many games where the Pentium G2120 was tested in and it had good FPS.

http://techreport.com/review/23662/amd-a10-5800k-and-a8-5600k-trinity-apus-reviewed/16
 
@oxford373 and blazorthon
With what you guys were talking about HTT in mind, would it be recommendable to disable HTT on Core i3's for games, even if they can take advantage of multiple threads? I bet this may depend on the game, but generally based on what you guys have seen? :)
 
[citation][nom]army_ant7[/nom]@oxford373 and blazorthonWith what you guys were talking about HTT in mind, would it be recommendable to disable HTT on Core i3's for games, even if they can take advantage of multiple threads? I bet this may depend on the game, but generally based on what you guys have seen?[/citation]

HTT should be left on for i3s AFAIK. Just look at the difference between the Pentium G2120 and the the i3-3225 in the picture that I posted above. HTT can make a significant difference in real-world performance for Intel's dual-core CPUs. Its not until you get to and beyond the quad core variants that Intel's HTT starts to not really make a difference according to what I've seen.

Now if we were to discuss HTT on Intel's now-ancient HTT-capable Netburst CPUs, then it might be a different story.
 
Thanks for the reply blaz. What got me thinking this was what oxford mentioned here:
Could it be just because it (i3-3225) has a higher clockrate (than the Pentium G2120)? Though the difference in performance appears to have not just come from just a boost in clockrate. (I didn't really check any specific benchmarks to see what the difference in performance really is.)

Though like you said here which I initially failed to pay particular attention to:
I guess you were saying Metro 2033 may be an exception which very well may be true.

 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom