billcat479 :
I have to agree about this screwy article. I sat and watched the differences and there is no way anyone would notice it while playing a game. And the difference seemed more with shading than how sharp the image is.
About the site being biased that is a no shocker as it went all the way back to the Pent 4 days and thrashing AMD because they were still using the equiv speed and Intel was still recomened over AMD at the time AMD had a faster cpu EXCEPT when it came to rambus systems that had a higher memory and was good at a few tests while AMD beat the rest.
And back then all the Intel inside ads on the site was a real giveaway about who was paying who for those articles to come on like that. I stopped coming to this site and Anands site as well. That site was totally in Intel camp.
It looked like things changed here so I came back but I don't really trust it all that much because of the way the two cpu camps are tested. Even knowing one is faster than the other in some apps is pretty much the norm to make one look better over the other.
While no one made much of it they sure have been raising a stink when the video cards and tweaked benchmark programs tried that trick to give the illusion of better performance.
Even though AMD can not beat Intel they can still come up close on many tests and even pass intel on others but you have to go out and look at a lot of sites before you start to see what is going wrong.
In the end the reviews are more about high profile articles to attract more people because this is how these sites make their money. When this site did a review on Intels newest cpu they made the mistake of putting AMD in there instead of a new Intel against the older one. The new Intel cpu test ran software that it seemed to be designed to work with and AMD in that review beat Intel's older one in most of the tests. I always thought that one was funny because in that article they rarely mentioned AMD's performance in that test even though it did pretty damm good. I was at best confused because prev. tests ran different benchmarks where the AMD got trashed by the older Intel cpu and then it beats it when they throw in that new Intel cpu. With the type of tests ran the picture changed quite a bit about where AMD's strengths were in their new design. I'm not defending AMD here, I'm making a point of a long standing drawback about review sites that show what people want to see so they visit more and the sites make more money. That is the bottom line.
And like it or not they played a small hand in AMD's sales because if the whole picture was shown back then AMD could be in a better position now but these sites were MHZ crazy back then and the super high clock rates was a lot more cool against a cpu that was clocked a lot lower but was faster but that isn't how it was played out when it first hit the market or for many months after. It wasn't until it was super obvious that AMD at the time had a better cpu and Intel had a egg cooking joke that had to slow down a lot or it would die from heat problems. In a way Intel had to take a better look at AMD's cpu because they took the same path and made it better over time because they have the resources once committed to the slower but faster AMD design.
And the rest is history.
Being easily noticeable is not a prerequisite for something to be a problem. I don't really care whether or not Tom's recommended P4s over Athlon 64s unless you have a point to make. Ever since then, Intel has had faster CPUs. The only time that AMD came close again was with Phenom II versus Core 2. Since that time, Intel has only improved and AMD has, overall, declined. I don't care about the brand name of the CPU or really any other part in my system a whole lot. I would buy Intel for a gaming machine right now because Intel's i5 CPUs are the best for that due to a huge diminishing returns on gaming performance increases by going over an i5 since all performance increases going over the i5s is simply the core and thread count increasing along with minor cache increases.
AMD is not for gaming anymore, or at least not high end gaming. Some of AMD's CPUs are okay up to 1080p gaming. All of AMD's *high end* CPUs only have high end performance for the most highly threaded work that uses either 6 (Phenom II x6 and FX-6xxx) or 8 (FX-8xxx) threads and utilizes them well. AMD doesn't compete with Intel in performance per core/thread anymore. If that changes, then I'll gladly use an AMD processor in my system. However, it probably won't because that would mean that AMD needs to not only make up a huge amount of lost ground, but also then beat Intel significantly. That would require something like doubling FX's IPC while decreasing power usage. Is that even possible right now?
Well, only if AMD can do another die shrink to 22nm and modifies the design and architecture greatly. I can see Piledriver cores matching Nehalem for IPC and almost matching Sandy Bridge for power efficiency because I've done the math and it's possible. Whether or not AMD managed to do what I thought they should do or did even better is up for discussion, but won't get anywhere and claiming to know how much faster Piledriver will be than anything is just wrong, unless you are inside AMD or something like that. No amount of leaks will tell for certain what something will do. If I remember correctly, Bulldozer was supposed to be as fast as or faster than Sandy Bridge according to many of it's *leaks*, but it doesn't even come close.
AMD isn't being shown in an unfair light here. Sandy Bridge is about 50% faster than Bulldozer at the same clock speed. Phenom II isn't much better than FX. and doesn't match FX's clock frequencies, so the two are roughly equal and also roughly equal with Core 2.
As for an Intel CPU beating an AMD CPU and then losing later on? Are you referring to something like a Core 2 Duo to AMD quad core comparison, because if so, then the test that was performed more recently was probably more multi-threaded than the previous test. If the previous test used only one or two threads than it makes sense for the Intel to win if it's clock speed isn't significantly lower than the AMD and then lose if the test becomes quad threaded. A lot of games are either quad threaded or becoming quad threaded now, so at the low/mid end, AMD and Intel are still competing because of Intel's lower core counts at the low end that more or less negate the performance per core difference.
For example, an FX-4100 and i3-2120 trade places as the winner in different games, with the FX winning somewhat in some quad threaded games and the i3 winning somewhat in in lightly threaded games and quad threaded games that don't spread the load around evenly (one or two heavy threads, the rest are light threads. This hurts AMD, but helps Intel's i3s because the i3s have Hyper-Threading, so they have two heavily and two lightly performing threads. The heavies are the physical threads and the lights are the logical threads).
When you actually look for reasons for Intel to win and for reasons for AMD to win, it's a lot easier to find the truth in the matter because you're not being biased by wanting a specific brand to win. AMD's best gaming processor from the FXs is the FX-4100 and their best Phenom II is the 960T (if overclocked because it should overclock better than the Deneb quad cores do). Given a choice between the two, the 960T is the obvious winner. Even if it can't be unlocked to six cores, it still has more IPC and should be able to have about the same clock frequencies as the FX processors. Whether it will be more or less power efficient, I'm not certain. However, it would still be slower than Intel's Nehalem quad core CPUs and they are significantly slower than Sandy Bridge CPUs which are somewhat slower than the upcoming Ivy Bridge CPUs. We already pretty much know how the Ivy's perform (not enough info on low/mid end overclocking just yet) at stock, but we don't really know how well Piledriver cores will stack up.
As of right now, it's AMD that is faster (for the price) only in some apps. Intel rules most programs so long as they aren't more than quad threaded. Most non-workstation and server programs are single, dual, or quad threaded, with single and dual still being the most common, but quad is catching up little by little. This is why SMD is losing... AMD's many slow cores approach due to their apparent inability to make something significantly faster than Phenom II does not go well with software that only uses 70% or less of the cores in the AMD processors. A six or eight core FX is roughly equal to a quad core FX at the same frequency in lightly and quad threaded software and thus is just wasting money and electricity.
i3s and FX give nearly identical performance in gaming (from a visual standpoint. 10-15% faster in either way will not be a very noticeable difference unless there is something else seriously wrong with the machine), but FXs do it at about 75-150% more power than the i3s do. Even when the FXs are cheaper, it only takes two to three years to recoup your losses if your computer is on for about four to six hours a day. Even faster if it's on more often than that. You don't even need to be using it because unless you outright shut it down, the FX uses a lot more power (as proven by every site who tested power usage. They can't all be biased like you say they are so much that they lie about power usage) than the i3s. If the machine is on more like 18 hours a day to 24/7, then you will recoup your losses on slightly higher upfront costs several times over during the lifetime of the machine unless it manages to die within a few months. I get my numbers from an AMD fan to, so no, I'm not using bad numbers.
Face it, AMD is no good for high end gaming and is more expensive over a few years than Intel for low/mid end gaming. AMD's low/mid end CPUs use more power than Intel's high end i5s and i7 quad core CPUs! AMD is really only very useful for low/mid end highly threaded environments (for desktops, AMD is really killing Intel in the mobile markets because of Intel neglecting them compared to the desktop market) or very low end PCs right now. Hopefully, Piledriver/Trinity will change that, but it might not.
My Phenom II x6 1090T was a far better purchase than any Intel processor with a similar cost for my work machine because of it's highly threaded performance. However, I realize that for gaming and such, it would be slower than an i3 for lightly threaded performance and not to much faster than an i3 for quad threaded performance. Considering my 25% overclock to just over 4GHz, it would beat the i3s in quad threaded performance significantly, but not in power efficient since it already uses about twice as much power at stock as the i3s use. However, for highly threaded performance, it is pretty fast. It hangs with the FX-8100 in highly threaded performance and they are right behind the i7s for highly threaded performance.
Facebook
Twitter
Instagram