ITT: We propose our own Socket AM3+ SKUs

[Phynaz]

FX4 is of i3 level on applications

Lol wut?


http://www.anandtech.com/bench/product/1273?vs=1197

Did you know that knowingly posting false information is against the forum's terms of service?

 

[Phynaz]

Before getting further in your usual thread crapping of AMD related thread get a look at the slide i posted and at Hardware.fr site, thoses numbers are documented, now please find an Intel thread if you want, there s enough of them by here, and go post there rather than trolling all the way whenever you read "AMD".

You're one slide outweighs his five links?

I don't think you understand the concept of "proof".

 

[Abwx]

Lol wut?
http://www.anandtech.com/bench/product/1273?vs=1197

Did you know that knowingly posting false information is against the forum's terms of service?

Certainly the i3 is better in games but in applications the FX43xx are largely competitive.

Applications :

http://www.hardware.fr/articles/913-7/cpu-performances-applicatives.html

Games :

http://www.hardware.fr/articles/913-8/cpu-performances-jeux-3d.html

There s the average of all thoses benches at the bottom, so who is spreading false information.?.

And i m generous, because of the whinings of some fans i m using a biaised rule with 10 applications being weighted as equal quantity as a single application that is gaming, indeed since a few times it look that we are in console dedicated forum...

You're one slide outweighs his five links?

I don't think you understand the concept of "proof".

Proof, you mean urban legends wich are your usual arguments.?.

I posted hard numbers, now you can always ask for some others "proofs" that suit your bias, there s a lot of incompetent sites here and there if you want to get some convenient numbers.

 

[Abwx]

Certainly the i3 is better in games but in applications the FX43xx are largely competitive.

Indeed i m also relying to urban legends but i had the old hardware.fr game suite as excuse, more recent results are different but they published only a few, actualy i stand corrected by Phynaz himself since the link he provided show that even in games the FX4350 is as good as the higher clocked i3...

http://www.anandtech.com/bench/product/1273?vs=1197

Edit : Here one can get a MB and a 4350 for the price of this i3..

 

[el etro]

This Hardware.FR is a damn good review site, awesome tests.

 

[BSim500]

Before getting further in your usual thread crapping

Your fake accusations are fooling no-one, especially when you're the one who repeatedly does exactly what you accuse others of on Intel threads on exactly the same subject without anyone else stooping to your level of spamming fake "thread crapping" accusations... (And you certainly have more experience on that subject anyway). I also didn't realize that Anandtech rules have changed and now require everyone to "join a CPU manufacturer gang" and only post on topics with your "tribe's" brand name mentioned in the title...

You're one slide outweighs his five links?

Abwx is obsessed with hardware.fr simply because they measure power consumption in a different way and it agrees with what he wants to hear. Click on hardware.fr's 220v from his own link) and it's suddenly back up to 76w (which is still 20% lower than everyone else measuring the same chip in the same way partly because they only use an obscure "Fritz chess" benchmark which loads the CPU by 20% less than the universally used Prime). Add 20% back as if they used Prime (or test it the same with prime) and guess what - power consumption is back up to +90w in line with everyone else's which he swears blind are all "wrong"... Meanwhile for Intel CPU's Abwx will quite happily quote AVX Prime measurements at the wall as if hardware.fr & Fritz Chess don't exist... Using his figures and "creative accounting", you could argue that Intel's 84w i5's "only really draw 49w. It's the truth, hardware.fr said so". Every other site on the net who measures it in a more standardized way (including Anandtech's reviewers) is declared "incompetent" and anyone who dares to point out the obvious discrepancy of "hardware.fr vs the world" due to their choice of software not fully loading the CPU is "thread crapping" and spreading "urban legends"...

Enigmoid picked up on his endless "apples & oranges" power consumption figures here too (as have several others in the past). It's something we've all seen before and has been done to death elsewhere.

 

[ShintaiDK]

I guess my 4670 is 40W TDP then with his wierd "math". But again, Nostra post more valid posts than he does. So lets stay with it as a 84W part. No need to go wrong in town due to misinformation.

 

[SPBHM]

Your link is with Prime 95 and a 100W delta at the main will yield close to 80W at the CPU level, exactly the number i posted above for P95, also under Prime the FX is stressed not only with AVX but also FMA.

The power figure i got are from hardware/fr review of the 6350/4350, the 62W are the CPU TDP under Fritchess.

The overclocking tests use Prime 95 while the regular compsumption is measured with Fritzchess :

http://www.hardware.fr/articles/899-3/influence-turbo-undervolting-overclocking.html

Column ATX 12V is the power drain measured before the VRMs, this has to be factored by 0.9 for the CPU actual TDP, the stock settings is the first tested with stock voltage, at 88.8W before the VRMs, all figures are with Prime 95, including when with stock settings.

I feel like this discussion is off topic but, as always you should take as many sources in consideration as possible, and if you add hardware.fr to all the others we have, it's clear that the 4350 is using a lot of power for how it performs; it's not really comparable to the i3.



looking at hardware.fr their atx12 fritzchess thing you like
http://www.hardware.fr/articles/901-2/consommation-efficacite-energetique.html

FX 4350 = 68.4W
i3 4340 = 34.8W

and the i3 I linked for performance comparison was a 35W model, while the 4340 is a 54W model,

still, going by their test, the 4340 is using half the power on fritz chess atx12
it's a little bit faster for their applicative average, clearly faster for gaming...
the cheaper 4150 should deliver basically the same performance as the 4340.

only really aggressive pricing can save the 4350, and the most basic chipsets for AM3+ (like 760G/SB710) are not as good as H81.

 

[Abwx]

Abwx is obsessed with hardware.fr simply because they measure power consumption in a different way and it agrees with what he wants to hear.

They measure power at the wall and within the plateform, as such they isolate the CPU and plateform powers, now find another site that make such extensive measurements, indeed i prefer professionaly behaving sites.

Click on hardware.fr's 220v from his own link) and it's suddenly back up to 76w (which is still 20% lower than everyone else measuring the same chip in the same way partly because they only use an obscure "Fritz chess" benchmark which loads the CPU by 20% less than the universally used Prime). Add 20% back as if they used Prime (or test it the same with prime) and guess what - power consumption is back up to +90w in line with everyone else's which he swears blind are all "wrong"...

You are ignorant and use ignorance as argument, they are clear on their protocol, Fritzchess bench load heavily the CPU mainly, so they can isolate the CPU comsumption in ST and MT, they use Prime 95 in overclocking tests to check comsumption and stability, so we have the peak comsumption at stock settings and with their users settings since they start their ocking test by stocks conditions with Prime 95.

Insulting other members is not allowed
Markfw900

Now if you are ignorant of Fritzchess go download the bench and do some tests but just stop putting the blame of your clulessness on Hardware.fr.

Meanwhile for Intel CPU's Abwx will quite happily quote AVX Prime measurements at the wall as if hardware.fr & Fritz Chess don't exist...

That s a lie, find the quote where i m saying so, but it s usual with you, hence you are a troll since you have to deseperatly rely in lies continously, i rarely saw someone being as liar when it comes to create false statements.

Using his figures and "creative accounting", you could argue that Intel's 84w i5's "only really draw 49w.

In Intel reviews they do the same measurements, go check rather than continously thread crapping AMD related threads.

 

[TurtleCrusher]

When I've left the FX-8350 at 2.8Ghz in CCC and gamed I had no idea I did that until I went to turn it down and it wasn't at 4.2ghz. Performance was no different in feeling than at 4.2ghz. I'm sure at 2.8ghz and 1.2v or so it's probably hitting 65-70w.

FX-8300E, 2.8GHZ, 3.4GHZ turbo. Disable 2MB of the l3 if need be to hit 55-65w.

 

[Abwx]

I feel like this discussion is off topic but, as always you should take as many sources in consideration as possible, and if you add hardware.fr to all the others we have, it's clear that the 4350 is using a lot of power for how it performs; it's not really comparable to the i3.
looking at hardware.fr their atx12 fritzchess thing you like
http://www.hardware.fr/articles/901-2/consommation-efficacite-energetique.html

FX 4350 = 68.4W
i3 4340 = 34.8W

and the i3 I linked for performance comparison was a 35W model, while the 4340 is a 54W model,

still, going by their test, the 4340 is using half the power on fritz chess atx12
it's a little bit faster for their applicative average, clearly faster for gaming...
the cheaper 4150 should deliver basically the same performance as the 4340.

only really aggressive pricing can save the 4350, and the most basic chipsets for AM3+ (like 760G/SB710) are not as good as H81.

Granted the i3 use less power but that s not hyge amounts overall, in this test the performance is 7497 nd/s for the FX4350, 7240 nd/s for the i3 4340 and 6857 nd/s for the i3 4130, clearly the i3 has a better efficency despite the slightest better score of the FX, the i3 4150 would be of FX4300 level and difference are not big within this segment anyway, on the other hand the difference in price favour the FX largely.

About the plateform power comsumption Hardware.fr, like many other, use an expensive and power hungry MB since they overclock better but most users of such CPUs will get more basic boards, what matters is SATA 3 and USB 3, these exists even in 760G/785G based MBs, or more systematicaly on 960G.

 

[Phynaz]

but it s usual with you, hence you are a troll since you have to deseperatly rely in lies continously, i rarely saw someone being as liar when it comes to create false statements.


In Intel reviews they do the same measurements, go check rather than continously thread crapping AMD related threads.

Your constant name calling and insults speak louder than any factual point you are attempting to make.

 

[Abwx]

I guess my 4670 is 40W TDP then with his wierd "math". But again, Nostra post more valid posts than he does. So lets stay with it as a 84W part. No need to go wrong in town due to misinformation.

There s a review of this CPU at Hardware.fr, you can go there to check the numbers rather than using ignorance as an argument to make convenient and trollish assumptions with made up numbers, look like it s an habit with some brand loyalists, but whatever, the purpose is not accuracy , it s thread crappind ad nauseam of whatever is related to AMD.

I guess my 4670 is 40W TDP then with his wierd "math". But again, Nostra post more valid posts than he does. So lets stay with it as a 84W part. No need to go wrong in town due to misinformation.

There s a review of this CPU at Hardware.fr, you can go there to check the numbers rather than using ignorance as an argument to make convenient and trollish assumptions with made up numbers, look like it s an habit with some brand loyalists, but whatever, the purpose is not accuracy , it s thread crappind ad nauseam...

When I've left the FX-8350 at 2.8Ghz in CCC and gamed I had no idea I did that until I went to turn it down and it wasn't at 4.2ghz. Performance was no different in feeling than at 4.2ghz. I'm sure at 2.8ghz and 1.2v or so it's probably hitting 65-70w.

FX-8300E, 2.8GHZ, 3.4GHZ turbo. Disable 2MB of the l3 if need be to hit 55-65w.

The reviews i linked allow to do an accurate estimation of the TDPs and voltage required at a given frequency when they do ocking test with close steps like the one of the FX4350, an FX83xx at 2.8 should yield 56-57W with Prime 95, about 46-47W with Fritzchess or Cinebench, this with the usual voltage margin used by AMD or Intel.

 

[SPBHM]

Granted the i3 use less power but that s not hyge amounts overall, in this test the performance is 7497 nd/s for the FX4350, 7240 nd/s for the i3 4340 and 6857 nd/s for the i3 4130, clearly the i3 has a better efficency despite the slightest better score of the FX, the i3 4150 would be of FX4300 level and difference are not big within this segment anyway, on the other hand the difference in price favour the FX largely.

About the plateform power comsumption Hardware.fr, like many other, use an expensive and power hungry MB since they overclock better but most users of such CPUs will get more basic boards, what matters is SATA 3 and USB 3, these exists even in 760G/785G based MBs, or more systematicaly on 960G.

well, their system power at the wall is 135 vs 78, that's significant when you combine it with not as good performance, sure 4350 power usage is not outside of what you could expect/accept for a desktop CPU, but it can use more power than some way faster CPUs

the 4350 seems to have low availability compared to the i3s, so it's difficult to comment on pricing, but considering gaming and popular programs the i3 4150 is clearly superior, and if you add power efficiency... as I said the FX would need a big price advantage to have any appeal.

960G!?

760G (specifically the SB they use, SB710) does not support USB 3.0 and Sata III, to add USB/sataIII you have to use pci express external chips, which can limit performance and increase cost compared to the h81 solution.

 

[cbn]

960G!?

He must mean the 980G chipset.

 

[Abwx]

Your constant name calling and insults speak louder than any factual point you are attempting to make.

We recognize the trolls and other thread crappers by their own posts, there s no technical datas, Watts, performance numbers, in contrast look at my posts, theses are about numbers and how they were measured, or eventualy estimated, but surely not filled with constant ad hominems and willfull misquotes when not deliberate forged quotes, there s a lot of people in this thread whose interest is obviously not technical discussion but constant bashing of a brand, as said their posts speak for themselves...

 

[BSim500]

You are ignorant and use ignorance as argument, they are clear on their protocol, Fritzchess bench load heavily the CPU mainly, so they can isolate the CPU comsumption in ST and MT, they use Prime 95 in overclocking tests to check comsumption and stability,

From your own link on the chart of your own figures which clearly states "Fritz Chess Benchmark 4.3" and not "Prime" on the power consumption tests:-

"Consumption measures are not to be taken as absolute maximum values ​​but rather typical of a heavy load, since specialized software in the processor stress such as Prime95 can consume about 20% more"
https://translate.googleusercontent...99-2/consommation-efficacite-energetique.html

This is why hardware.fr's figures are consistently lower than everyone else's even comparing "at the mains" draw and ignoring 12v vs mains measurements - it's not because hardware.fr are "more accurate" but simply because they use software which puts a lighter load on the CPU (the equivalent of 80% Prime by their own admission) vs everyone else's industry standard Prime95 load tests. Which like I said is "apples and oranges" and will remain so no matter how much you stamp your feet and spew out all the petty insults you can think of every time someone points out the obvious. Further subtracting another arbitrary 10% for "power supply losses" is also great for a component isolation test bench but not very useful for the average person who um, uses an 80-90% efficiency power supply to power their computer and may actually experience even greater losses in practise than what hardware.fr are attempting to "compensate" with an arbitrary "guesstimate".

Protest as much as you like until you are red in the face - real-world maximum power consumption of CPU's is usually higher than hardware.fr's skewed figures as figures from pretty much every other benchmarking / review site proves (not to mention anyone with their own Kill-A-Watt). When 19 sites say one thing and 1 certain other consistently comes up with statistical outliers, that doesn't mean that 1 site is "right" and everyone else "trolling", it's simply a stark demonstration of your "confirmation bias".

That s a lie, find the quote where i m saying so

We recognize the trolls and other thread crappers by their own posts

Here's one of many. So no it isn't a lie. And as for "thread crappers", we sure do recognize them alright... When quoting AMD power consumption figures you cherry pick hardware.fr's "80% equivalent Prime load and subtract another 10% on top" ones and when quoting Intel's you regularly pick the highest over-volted AVX Prime ones as and when it suits you. You claim i5 S chips are "77w" and yet "K" chips on hardware.fr are only 56w. Then you claim everyone except you is doing it! There are many other more glaring examples of your chronic self-contradictions and brand bias related double standards, but it's New Year's Day and I really can't be bothered to respond to your "how dare you defy me!" overly-defensive style of posting or general "attitude" anymore, so I'll simply end with a polite "Happy New Year".

 

[Abwx]

well, their system power at the wall is 135 vs 78, that's significant when you combine it with not as good performance, sure 4350 power usage is not outside of what you could expect/accept for a desktop CPU, but it can use more power than some way faster CPUs

the 4350 seems to have low availability compared to the i3s, so it's difficult to comment on pricing, but considering gaming and popular programs the i3 4150 is clearly superior, and if you add power efficiency... as I said the FX would need a big price advantage to have any appeal.

960G!?

760G (specifically the SB they use, SB710) does not support USB 3.0 and Sata III, to add USB/sataIII you have to use pci express external chips, which can limit performance and increase cost compared to the h81 solution.

Sorry i messed the chipset name..

SATA 3, USB3 will use third part controlers but perfs are good as showed by a member in another thread, a FX 43xx is highly variable here, as low as 72€ in some european countries but hardly less than 80€ in France , an equivalent i3 start at 120€, in contrast MBs with both USB3 and Sata 3 can be found at 48€, essentialy you have the CPU + MB for the price of the lower priced i3s, those lowly priced MBs wont consume the amount of a 990FX based MB, wich was the plateform used by HFR.

 

[Abwx]

"Consumption measures are not to be taken as absolute maximum values ​​but rather typical of a heavy load, since specialized software in the processor stress such as Prime95 can consume about 20% more"
https://translate.googleusercontent...99-2/consommation-efficacite-energetique.html

What they are saying is that Prime 95 use about 20% more power on average than their power comsumption tests with Fritzchess bench, this is indeed clearly visible when comparing the power comsumption numbers they have with Fritz and with Prime 95, for the FX 4350 at the ATX 12V connector the figures are, at stock settings, 68.4W and 88.8W respectively, this is clearly displayed here for power under Fritzchess both ST and MT :

http://www.hardware.fr/articles/899-2/consommation-efficacite-energetique.html

And here for the results under Prime 95 :

http://www.hardware.fr/articles/899-3/influence-turbo-undervolting-overclocking.html

All the numbers are in the review, it s just you that are ignorant of what they are talking about, as proved by the useless post i m answering to.

 

[cbn]

SATA 3, USB3 will use third part controlers

I might be getting confused about what board chipset series you are referring to, but I have seen AM3+ boards with integrated graphics and native SATA III (6 Gbps) together.

One of them was a 880G board with SB 850.

But yes, the usb 3.0 needs to be third party at this time.

 

[ShintaiDK]

I might be getting confused about what board chipset series you are referring to, but I have seen AM3+ boards with integrated graphics and native SATA III (6 Gbps) together.

One of them was a 880G board with SB 850.

But yes, the usb 3.0 needs to be third party at this time.

Without investigating if there is such a SKU or not.

There is nothing hindering them to use a newer southbridge. The A-Link Express interface is just another PCIe interface like DMI.

 

[TurtleCrusher]

We recognize the trolls and other thread crappers by their own posts, there s no technical datas, Watts, performance numbers, in contrast look at my posts, theses are about numbers and how they were measured, or eventualy estimated, but surely not filled with constant ad hominems and willfull misquotes when not deliberate forged quotes, there s a lot of people in this thread whose interest is obviously not technical discussion but constant bashing of a brand, as said their posts speak for themselves...

The difference between enthusiast and fanatic is strikingly clear. Informed speculation and thought provoking discussions is what makes an enthusiast community a good place to discuss. Those who post rampant misinformation and parrot talking points are what ruins a community such as this.

 

[Phynaz]

We recognize the trolls and other thread crappers by their own posts, there s no technical datas, Watts, performance numbers, in contrast look at my posts, theses are about numbers and how they were measured, or eventualy estimated, but surely not filled with constant ad hominems and willfull misquotes when not deliberate forged quotes, there s a lot of people in this thread whose interest is obviously not technical discussion but constant bashing of a brand, as said their posts speak for themselves...

The above is no excuse for telling people they are ignorant trolls. Mark acted on my report, maybe you should take his warning seriously, as I will continue to report every post that you insult somebody. Hopefully others will do the same.

 

[cbn]

Looking through the AM3+ Vishera line-up, all the chips (with the exception of FX-4300 which has 4MB L3 cache) have the full 8MB L3 cache:

http://www.anandtech.com/show/6396/the-vishera-review-amd-fx8350-fx8320-fx6300-and-fx4300-tested

http://www.anandtech.com/show/8427/amd-fx-8370e-cpu-review-vishera-95w

However, looking at the die shot wouldn't there be a lot of room for AMD to offer full octocores with just some (or even all) of the L3 cache disabled?

Based on what I am seeing below the chances of a defect being in that cache would be just as high as one of the cores.

 

[Shehriazad]

How about a 125/140W FX7590? xD

5GHZ 6core.

But then again...that would probably destroy their own productline.