Vishera Review Up - Anandtech

[inf64]

Why are you so obsessed with C11.5 performance? It's just one test and it may be a glitch. Look here:
http://www.guru3d.com/articles_pages/amd_fx_8350_processor_review,12.html
You see how C11.5 numbers can vary? Guru3d supposedly get as low as 3.34pts in this benchmark for FX6100,how's that possible? It's 43% lower than what you get,yet it's still in their charts for over a year... Let go of this one benchmark.

 

[Shakabutt]

Why are you so obsessed with C11.5 performance? It's just one test and it may be a glitch. Look here:
http://www.guru3d.com/articles_pages/amd_fx_8350_processor_review,12.html
You see how C11.5 numbers can vary? Guru3d supposedly get as low as 3.34pts in this benchmark for FX6100,how's that possible? It's 43% lower than what you get,yet it's still in their charts for over a year... Let go of this one benchmark.

How does that make it better, i dont think they(guru)even tested the old bulldozers after the hotfixes or on windows 8.

Their just pulling scores from a year ago,and its still stupid.

Finally,being lower than expected is one thing,but a whole .8 points higher is very unusual brother since you can get lower error scores bellow average but above? lol

 

[iCyborg]

I like that AT included i7 920 in the review, it's still a formidable CPU
FX8350 would be a sidegrade at best, even 3770K is not much of an upgrade for me. If it was on the same socket, I might be tempted, but a new MB is too much hassle...

 

[alexruiz]

Nobody here with a FX 6100 to confirm this ?

Real fishy

Every review besides the Anandtech one is showing some really weird scores for these processors, can anyone else confirm this ?

I'll make a run later and will give you numbers.
What is your setup (motherboard)?
Catalyst version?
RAM timings?

 

[Subyman]

A good showing from AMD. It is what they needed to actually be comparable at the mid range. This would be a great processor for those that heavily multitask. Too bad I don't need another computer, I'd love to have an AMD system after being with Intel since Conroe.

 

[Nemesis 1]

I'd like to see a gaming comparison of OC'ed FX-4300 vs Core i3

Didn't anyone do A highend 6 core intel in these reviews so as to put things in proper perspective.. My eyes not seeing what my ears hearing .. Ananda test confuse , If ananda had put a 2600k in the review it would be easily seen that sb to ib was a bigger step forward than was the amds step forward you can see this between 2500k and the 3750. IVB gains are greater than amds for the most part . Had ananda used a 2600k than it would have shown more clearly . But it would seem AT is giving AMD a leg up here which is fine AMD needs this type of help. Now intel will release a higher end IB 3800K 4 CORE 4 THREADS AND OF COURSE A HIGHER MODEL 8 THREADED CORE . tHAT WILL wipe OUT THE CLOCK ADVANTAGE WE ARE ALL SEEING IN THESE RESULTS . Intel can go alot higher on clocks and I for one expect an announcement in the next 2 weeks . Same crap differant day

 

[alexruiz]

Didn't anyone do A highend 6 core intel in these reviews so as to put things in proper perspective.. My eyes not seeing what my ears hearing .. Ananda test confuse , If ananda had put a 2600k in the review it would be easily seen that sb to ib was a bigger step forward than was the amds step forward you can see this between 2500k and the 3750. IVB gains are greater than amds for the most part . Had ananda used a 2600k than it would have shown more clearly . But it would seem AT is giving AMD a leg up here which is fine AMD needs this type of help. Now intel will release a higher end IB 3800K 4 CORE 4 THREADS AND OF COURSE A HIGHER MODEL 8 THREADED CORE . tHAT WILL wipe OUT THE CLOCK ADVANTAGE WE ARE ALL SEEING IN THESE RESULTS . Intel can go alot higher on clocks and I for one expect an announcement in the next 2 weeks . Same crap differant day

Several reviews used 2600k. Still good showing, check the numbers

 

[NTMBK]

Didn't anyone do A highend 6 core intel in these reviews so as to put things in proper perspective.. My eyes not seeing what my ears hearing .. Ananda test confuse , If ananda had put a 2600k in the review it would be easily seen that sb to ib was a bigger step forward than was the amds step forward you can see this between 2500k and the 3750. IVB gains are greater than amds for the most part . Had ananda used a 2600k than it would have shown more clearly . But it would seem AT is giving AMD a leg up here which is fine AMD needs this type of help. Now intel will release a higher end IB 3800K 4 CORE 4 THREADS AND OF COURSE A HIGHER MODEL 8 THREADED CORE . tHAT WILL wipe OUT THE CLOCK ADVANTAGE WE ARE ALL SEEING IN THESE RESULTS . Intel can go alot higher on clocks and I for one expect an announcement in the next 2 weeks . Same crap differant day

Woah, calm down man.

 

[inf64]

Apropos the comment that Ib->SB brought bigger gains,let's see some facts rather than fiction.
2600K to 3770K(a die shrink + design tweaks including frequency gains) brought 8.8% improvement in performance,as hardware.fr shows below.
On the other hand 8150 to 8350 (no shrink+design tweaks including freq. gains) brought ~15% improvement ,a lot more than what SB->IB got.

Similarly in games 3770K was 6% faster than 2600K while 8350 is 18% faster than 8150.

 

[beginner99]

Apropos the comment that Ib->SB brought bigger gains,let's see some facts rather than fiction.
2600K to 3770K(a die shrink + design tweaks including frequency gains) brought 8.8% improvement in performance,as hardware.fr shows below.
On the other hand 8150 to 8350 (no shrink+design tweaks including freq. gains) brought ~15% improvement ,a lot more than what SB->IB got.

Similarly in games 3770K was 6% faster than 2600K while 8350 is 18% faster than 8150.

relative comparisons have their issues. That 8.8% is about the same absolute terms considering SB had a about a 40%-50% lead in single threaded performance over BD. Plus it is in general much easier to improve something crappy than something good.

 

[Olikan]

sorry, i just had to...

 

[Shakabutt]

I'll make a run later and will give you numbers.
What is your setup (motherboard)?
Catalyst version?
RAM timings?

I have a

M5A78L-M LX motherboard

Catalyst 12.11 beta

Kingston Hyperx 1600 @ 9-9-9-24

 

[inf64]

@beginner99

I was just disproving the ridiculous claim,not saying anything about IB vs Vishera. But as you can see in application workloads,it's just mere 6% faster than 8350 and costs a lot more. For those not primarily gaming it's a nice alternative.

 

[exar333]

Apropos the comment that Ib->SB brought bigger gains,let's see some facts rather than fiction.
2600K to 3770K(a die shrink + design tweaks including frequency gains) brought 8.8% improvement in performance,as hardware.fr shows below.
On the other hand 8150 to 8350 (no shrink+design tweaks including freq. gains) brought ~15% improvement ,a lot more than what SB->IB got.

Similarly in games 3770K was 6% faster than 2600K while 8350 is 18% faster than 8150.

Google 'tick tock'. You completely missed the point. You should really look at gulf town to SB, which was a tweaked arch on 32nm. That was about 15% too. Intel die shrinks, like SB to IB, have traditionally been more like 5%. Look to Haswell to be 10-15% across the board and much more where AVX is related. That said, this was a decent improvement for AMD. It's easier to improve on a poor product than make material changes to a better one oftentimes.

 

[exar333]

Unfortunately for AMD, I don't think this opens up much more of a market. It's still pretty slow at gaming and power hungry, with much lower IPC. I think this really only tempts those who need a lot of MT performance that may have been on the fence before.

Edit: oh, and AMD 'loyalists' who would buy anything branded AMD in a box...

 

[AtenRa]

Found a video of some youtube guy cinebenching his fx 6100 @ 4 GHZ result : 4.8
exactly what im getting

http://www.youtube.com/watch?v=yf0hc3Y046U

And yet those sites get 4.8 AT STOCK.

Talk about misinformation sigh.

Edit : this is what im talking about

http://nl.hardware.info/reviews/331...0-fx-6300-getest-benchmarks-cpu-cinebench-115

http://www.legionhardware.com/articles_pages/amd_fx_8350_and_fx_6300,3.html

http://www.techspot.com/review/586-amd-fx-8350-fx-6300/page3.html
Almost identical review.Next time mix your reviews up before you pay for them AMD.

Makes you wonder how much they inflated those Piledriver scores.

So from what i see Anand's review seems to be spot on, while these other ones are like from the same bullshit cloth or something.

Yeap, the numbers at Cinebench 11.5 must be wrong, FX8150 gets close to 6.00 not 7.04.

 

[inf64]

If you look at SB scores in their review you will see them also higher than normal (what others get).

 

[Jovec]

Apropos the comment that Ib->SB brought bigger gains,let's see some facts rather than fiction.
2600K to 3770K(a die shrink + design tweaks including frequency gains) brought 8.8% improvement in performance,as hardware.fr shows below.
On the other hand 8150 to 8350 (no shrink+design tweaks including freq. gains) brought ~15% improvement ,a lot more than what SB->IB got.

Similarly in games 3770K was 6% faster than 2600K while 8350 is 18% faster than 8150.

Regarding 8.8% vs 15%:

To me it's not a direct comparison, and ultimately as consumers it doesn't matter to us, but AMD is running their CPUs essentially as fast as they can. The 8350 would be at 4.2, 4.4, or even 5.0 stock if AMD could manage it within their power and cooling envelope.

Intel clearly isn't, as they have different power and profit goals and the luxury of mind share and process tech. Give Intel AMD's die size/transistor count and power budget and they would absolutely crush AMD. For example, are there any 2500 or 2600 CPUs in the wild that won't operate with an extra 200MHz at stock volts? How about at 4.0Ghz with a slight voltage tweak? If Intel wants to take a hit to their profits, just lower prices and any AMD value proposition goes away. Or sell 6c/12t SB/IB at reasonable prices.

 

[inf64]

I was just disproving what one of the posters was claiming . He said that IB brought greater increase in performance over SB than Vishera brought over 8150. I just proved him wrong. I have no doubts that intel can run 3770K at 3.7Ghz or higher at stock if they wanted to . Why they actually don't is a whole other matter.

 

[Olikan]

woot, found something interesting...piledriver have the same problem as bulldozer, 20% ipc hit in CMT
http://www.extremetech.com/computin...ledriver-deliver-where-bulldozer-fell-short/2

 

[Ketchup]

I just bought my i5 system a month ago. After looking at the Anandtech review, I am not disappointed with my purchase.

Not in the least.

This is the best you can do AMD?

 

[Hubb1e]

Pretty disappointing Anandtech article. Not much detail about gaming on the FX chips. When you look at the 4300 and 6300 pricing they look pretty good against the Intel i3s for gaming. Figure in the locked clocks and dual cores of the i3s and you've got a win for AMD if you can ignore power consumption which isn't THAT big of a deal.

 

[inf64]

woot, found something interesting...piledriver have the same problem as bulldozer, 20% ipc hit in CMT
http://www.extremetech.com/computin...ledriver-deliver-where-bulldozer-fell-short/2

Of course it has the CMT perf. hit,it has the same shared decoder and same shared (single) FP unit. That's why they claim 80% performance of "traditional" dual core design in less die area. Sharing of the frontend and FP unit makes it less efficient versus designs with dedicated execution units(like Phenom/Core). But looking at the MTed performance those 4 FP units can now significantly outperform Thuban's 6 cores each of which has a full fp unit to itself. Problem is that still each integer core cannot fully utilize the whole FP unit when there is lightly threaded workload and Win7 scheduler didn't help. I wonder how much if at all Win8 will improve things.

 

[exar333]

It's pretty amazing when you take a step back. A year ago, everyone was expecting the 8C parts to compete with Intel 6C parts, 6C parts to compete with 4C intel parts, and 4C to dominate Intel 2C offerings. Now we consider the 6C intel parts untouchable, 8C competes with 4C (more or less) and 6/4C against Intel 2C. Nuts.

 

[inf64]

It's "8C" only if you fall for marketing slides. For all intents and purposes it's more like AMD's version of quad core with (better than SMT) multithreading. It has 4 FP "cores" that can run 8 threads and it competes well against 4 intel cores running 8 threads (in MT workloads). It still has lower singlethread performance so for those wanting an all around faster chip they are better off with i7. i5s are in tricky position since they have only a few ST workloads(games being dominant ones) that run better and at anything MTed they loose to Vishera,sometimes badly. Content creation and 3d rendering is now a clear win for AMD's "8C".