The Bulldozer Review: AMD FX-8150 Tested
by Anand Lal Shimpi on October 12, 2011 1:27 AM ESTCache and Memory Performance
I mentioned earlier that cache latencies are higher in order to accommodate the larger caches (8MB L2 + 8MB L3) as well as the high frequency design. We turned to our old friend cachemem to measure these latencies in clocks:
| Cache/Memory Latency Comparison | ||||||
| L1 | L2 | L3 | Main Memory | |||
| AMD FX-8150 (3.6GHz) | 4 | 21 | 65 | 195 | ||
| AMD Phenom II X4 975 BE (3.6GHz) | 3 | 15 | 59 | 182 | ||
| AMD Phenom II X6 1100T (3.3GHz) | 3 | 14 | 55 | 157 | ||
| Intel Core i5 2500K (3.3GHz) | 4 | 11 | 25 | 148 | ||
Cache latencies are up significantly across the board, which is to be expected given the increase in pipeline depth as well as cache size. But is Bulldozer able to overcome the increase through higher clocks? To find out we have to convert latency in clocks to latency in nanoseconds:
We disable turbo in order to get predictable clock speeds, which lets us accurately calculate memory latency in ns. The FX-8150 at 3.6GHz has a longer trip down memory lane than its predecessor, also at 3.6GHz. The higher latency caches play a role in this as they are necessary to help drive AMD's frequency up. What happens if we turn turbo on and peg the FX-8150 at 3.9GHz? Memory latency goes down. Bulldozer still isn't able to get to main memory as quickly as Sandy Bridge, but thanks to Turbo Core it's able to do so better than the outgoing Phenom II.
L3 access latency is effectively a wash compared to the Phenom II thanks to the higher clock speeds enabled by Turbo Core. Latencies haven't really improved though, and Bulldozer has a long way to go before it reaches Sandy Bridge access latencies.
Facebook
Twitter
RSS
430 Comments
View All Comments
Hector2 - Wednesday, October 12, 2011 - link
Ivy Bridge on 22nm (Sandy Bridge on steroids) is just around the corner. It's not going to be prettysuperccs - Wednesday, October 12, 2011 - link
If Ghz is BD's salvation, then why did you post some OC'd benchmarks? Cmon, really you are going to make us do it?chizow - Wednesday, October 12, 2011 - link
Anand, you said it yourself. AMD hasn't been competitive since 2006 and Core 2. Has there been a real choice in the CPU space for the last 5 years? Do we even need choice? The fact of the matter is Intel is only competing with *THEMSELVES* in the CPU space and this has been the case for the last 5 years. They need to keep dropping prices and releasing faster processors because that's the only way they are going to sell more CPUs, irrelevant of whatever drek AMD is releasing in their wake.I really wish tech writers and pseudo-economists would stop repeating the lie of a meme that "we need competition" in order for innovation to continue, when that's simply not the case.
bji - Wednesday, October 12, 2011 - link
Your comments are a gross oversimplification. There are many segments of the x86 market and AMD has been competitive in some of them over the past 5 years. Just because they haven't been competitive in your favorite segment (which I am guessing is high end) doesn't mean that they haven't been competitive at all.I bought an Atom board and a Zacate board and the Zacate was clearly and thoroughly superior. So AMD wins at the very low end.
At the low midrange I believe that AMD has always, except until maybe very recently, been the better choice; of course AMD has achieved this by basically making no money on their parts which obviously is not sustainable.
At the mid-midrange AMD has been sometimes competitive and sometimes not over the past 5 years. I personally bought a 6 core Phenom II last year around this time because the workload I care about - massively parallel software compiles - was heads and shoulders better on AMD than similarly priced Intel at the time.
Once you get into the high midrange and all segments of the high end, then I finally agree with your comments - AMD has not been competitive there for 5 years.
On the laptop front, AMD has mostly been uncompetitive in the time frame in question except where they have reduced price to an unsustainable (for them) level and stolen the occasional win. The only win for AMD in that space is, once again, their Atom equivalents which win in netbooks.
Your post is an oversimplification. However, if AMD doesn't pull a rabbit out of their hat, then for the next 5 years your oversimplification will end up being true.
Competition driving innovation is not a lie, and your appeal to people to stop repeating this truth will go unheeded so don't waste your breath.
chizow - Wednesday, October 12, 2011 - link
No, its not really a gross oversimplication, its just an accurate generalization born out repeatedly over the last 5 years.Intel has AMD beat in every single price-performance segment in the x86 market except in the rare cases where mGPUs come into play, like your Zacate example. Which is ironic in and of itself, given some of the asinine remarks AMD leadership has made in the past with regard to ULP products.
Just look at this low to mid-range line-up, the pricing segments have been steady for the last 4-5 years during Intel's dominance, the ONLY things that change are the SKU names and clockspeeds.
http://www.anandtech.com/show/4524/the-sandy-bridg...
This all occurs independent of any meaningful competition from AMD, as every single Intel option offers better performance for the same or slightly more money.
Even in multi-threaded apps Intel often achieves performance parity with FEWER cores and less TDP. Its embarassing really how Intel can even compete with AMD's 6 core processors with their *2* core processors even in the threaded apps you're referring to.
http://www.anandtech.com/show/3674/amds-sixcore-ph...
Tick Tock doesn't stop when AMD stumbles, nor does Intel's commitment to process fabrication leadership. Why? Because their investors and consumers demand innovation INDEPENDENT of what AMD does or does not do to compete. Competition does not drive innovation in the x86 desktop market and has not for the last 5 years, the bottom line does.
Al1458 - Wednesday, October 12, 2011 - link
AMD is the only viable alternative to Intel. If AMD should not get out of this slump, we could well see AMD leave the CPU market which would leave one CPU manufacturer, and we all know how restrained Intel is with its pricing...hoohoo - Wednesday, October 12, 2011 - link
I really want to see some server benchmarks: LAMP & IIS stack, DBMSs, HPC benchmarks.I do not think the real potential is being explored, (for FPU intensive code especially) a recompile is needed with a compiler that knows what a BD is.
Kyanzes - Wednesday, October 12, 2011 - link
Simply horrible. Surely there has to be some mistake.mapesdhs - Wednesday, October 12, 2011 - link
Sometimes I think it's a pity a company like IBM doesn't make a consumer chip, ie. a rival to Intel
which has enough cash and expertise to do the job properly, come up with something that can be
a good fit for both the consumer & server spaces. IBM certainly has plenty of CPU design experience
for servers, eg. the 4.25GHz 8-core Power7.
Ian.
FunBunny2 - Wednesday, October 12, 2011 - link
They did. Called the PPC. Nobody bought it.