AMD Vs Intel
-
@Pete-S said in AMD Vs Intel:
If AMD's where too slow then it was the wrong CPU for the application and it was a purchasing error, not an error on AMD's part.
This. AMD makes CPUs that are so monster, so fast, and has been since around 200 when the Athlon came out. Intel's top of the line was the Pentium III (the P4 never matched it) and AMD pulled ahead during that era, and then made 64bit and left Intel in the dust. AMD hasn't had a lack of "faster than anyone needs" processors since the 1990s.
-
@zachary715 said in AMD Vs Intel:
Yes you likely would save on Ryzen 3 vs i5, because they're not directly comparable products. A more fair comparison would be Ryzen 5 vs i5 or Ryzen 3 vs i3. I would think you'd still see some savings, but maybe not quite as significant.
AMD savings should be visible right now because Intel is having supply problems and so cost is going higher relative to the norm as people who require or demand Intel are getting charged a premium. For those able to use Genuine AMD (to use Intel's terminology) you get a bigger benefit right now than you normally do and desktop and laptop makers are adding AMDs to their lineups quickly so that they can keep machines on the market. That's really why they are pushing AMD, they've always known that it was a better value but that doesn't matter to them, what does matter is that they are able to get them and keep getting them for you so that you can keep buying more.
-
@dafyre said in AMD Vs Intel:
For your average business desktops, yeah, I'd go AMD if it saves you a chunk of money.
For video editing rigs or cad workstations, I'd do Intel CPUs + AMD GPU (oh, the irony).
Why are people using those types of apps, still using single-threaded applications? That's the only thing that Intel does better than AMD. With Intel's CPU shortages, the price difference is just making AMD look better and better. With Ryzen 3 out soon (we're seeing leaked performance reports now), the difference in performance even at that 1 special thing isn't going to be so much anymore. Plus, those apps are so old, or the companies that wrote them stopped any active development so long ago that you should probably be moving away from their use. We're talking at least 15 years since multi-core became the standard for x86 compatible hardware.
-
There's no advantage to AMD. To get an AMD with the same performance as a modern Intel, you're spending at or more on an AMD then.
-
@Obsolesce said in AMD Vs Intel:
There's no advantage to AMD. To get an AMD with the same performance as a modern Intel, you're spending at or more on an AMD then.
Really? https://www.tomshardware.com/reviews/cpu-hierarchy,4312.html Your claim seems rather fishy.
-
@Obsolesce said in AMD Vs Intel:
There's no advantage to AMD. To get an AMD with the same performance as a modern Intel, you're spending at or more on an AMD then.
Actually, according to Forbes testing, AMD commands a lead while being cheaper. In single threaded apps and games, Intel has a tiny lead (but not when cost is factored), but for business workloads, AMD was better in performance with a pretty massive price lead.
-
@Obsolesce said in AMD Vs Intel:
There's no advantage to AMD. To get an AMD with the same performance as a modern Intel, you're spending at or more on an AMD then.
AMD besting Intel has been the accepted knowledge for some time. There was a tiny blip in the early 2010s when AMD had no new products and Intel was going gangbusters and they finally managed to get a lead on them after a full decade of being so bad to be nearly a joke. (But remember, prior to 2000, Intel was awesome and AMD a joke, so it was a major reversal as it was.) But not since the Pentium IIIS had Intel had anything impressive until around 2010.
AMD was out of the game for only a couple of years and has been the leader when real world cost is considered, for most workloads, for several years again now, and nearly all of the late 32bit and all of the 64bit era.
-
@scottalanmiller said in AMD Vs Intel:
@Obsolesce said in AMD Vs Intel:
There's no advantage to AMD. To get an AMD with the same performance as a modern Intel, you're spending at or more on an AMD then.
Actually, according to Forbes testing, AMD commands a lead while being cheaper. In single threaded apps and games, Intel has a tiny lead (but not when cost is factored), but for business workloads, AMD was better in performance with a pretty massive price lead.
Maybe it's just the ones I compared. Example, the i7 7700K. An equal performing AMD costs the same or more.
-
@Obsolesce said in AMD Vs Intel:
@scottalanmiller said in AMD Vs Intel:
@Obsolesce said in AMD Vs Intel:
There's no advantage to AMD. To get an AMD with the same performance as a modern Intel, you're spending at or more on an AMD then.
Actually, according to Forbes testing, AMD commands a lead while being cheaper. In single threaded apps and games, Intel has a tiny lead (but not when cost is factored), but for business workloads, AMD was better in performance with a pretty massive price lead.
Maybe it's just the ones I compared. Example, the i7 7700K. An equal performing AMD costs the same or more.
Check the tests above. The 8700 vs the R7... the R7 crushed the 8700 even in single threaded, let alone multi-threaded, performance. And did it at $225 compared to like $360. It's not just faster and cheaper, it is a huge amount faster and incredibly cheaper. That's like 60% the cost, for noticeably more performance in nearly any use case.
-
@scottalanmiller said in AMD Vs Intel:
@Obsolesce said in AMD Vs Intel:
@scottalanmiller said in AMD Vs Intel:
@Obsolesce said in AMD Vs Intel:
There's no advantage to AMD. To get an AMD with the same performance as a modern Intel, you're spending at or more on an AMD then.
Actually, according to Forbes testing, AMD commands a lead while being cheaper. In single threaded apps and games, Intel has a tiny lead (but not when cost is factored), but for business workloads, AMD was better in performance with a pretty massive price lead.
Maybe it's just the ones I compared. Example, the i7 7700K. An equal performing AMD costs the same or more.
Check the tests above. The 8700 vs the R7... the R7 crushed the 8700 even in single threaded, let alone multi-threaded, performance. And did it at $225 compared to like $360. It's not just faster and cheaper, it is a huge amount faster and incredibly cheaper. That's like 60% the cost, for noticeably more performance in nearly any use case.
Why does the much older, less core, less threads, i7 outperform that later and bigger better R7?
-
@Obsolesce said in AMD Vs Intel:
Why does the much older, less core, less threads, i7 outperform that later and bigger better R7?
Where do you see that? I don't even see it being tested. It's a "quad core" test specifically designed to not show performance, but to make for a useless report. Check overall speed, not a false "if we cripple it and don't actually test it" speed.
So from that test, it sure looks like the AMD is blowing the Intel away. Not just from the numbers you showed (those numbers are produced from 33% of the Intel, but only 25% of the AMD), but also from the fact that people are making these kinds of test just to make false numbers to try to make the Intel look competitive!
That's how slow the Intel is, that this kind of test even exists.
-
After Ryzen, everything AMD released is good. and worth having.
-
@scottalanmiller said in AMD Vs Intel:
@Obsolesce said in AMD Vs Intel:
Why does the much older, less core, less threads, i7 outperform that later and bigger better R7?
Where do you see that? I don't even see it being tested. It's a "quad core" test specifically designed to not show performance, but to make for a useless report. Check overall speed, not a false "if we cripple it and don't actually test it" speed.
So from that test, it sure looks like the AMD is blowing the Intel away. Not just from the numbers you showed (those numbers are produced from 33% of the Intel, but only 25% of the AMD), but also from the fact that people are making these kinds of test just to make false numbers to try to make the Intel look competitive!
That's how slow the Intel is, that this kind of test even exists.
Oh so fake tests.
-
Testing single threaded or limited threaded performance of CPUs is useful only if we aren't talking about normal workloads and you know that your software can't use any modern CPU properly. In which case, none of these processors make the least sense, nor were they set up correctly even for the processors selected. The test is total BS.
If testing single or 6 or fewer thread limits, you turn off HT. They didn't, these are bogus even on just measuring an individual processor.
But the test is designed around 1990s video game workloads, not modern desktops. In the modern office, we expect people to run web browsers, office apps, and such. Even if they still run MineSweeper and some single threaded app, it's just one of many workloads. These tests are designed to ignore that.
The large number of AMD threads is because it is designed around real world workloads, not arbitrary tests. Sure, at some point, no one needs more threads, but for most of us, that's what we need the most. We have lots of open tabs, windows, and apps. Each one uses one or more threads, plus the OS would like a few of its own if they are available. I could use over 20 threads, right now, just for my browsers let alone anything else. A 32 thread processor would not be wasted for me or any of my staff, and we are light IT users.
If you have people running more intensive workloads, that need normally just increases. I'm not saying that procs need to go thread crazy, I'm just saying that testing them in a limited fashion reflects inversely on how they will perform in real desktop workloads.
-
This is a chart that shows CPU value, i.e. performance per $.
It includes CPUs that you can currently buy, so there are older models as well.It's Passmark's CPU benchmark which is just a bunch of different CPU operations like compression, encryption, sorting, floating point math, integer math etc.
It's pretty safe to say AMD is the best value in general.
-
@Obsolesce said in AMD Vs Intel:
@scottalanmiller said in AMD Vs Intel:
@Obsolesce said in AMD Vs Intel:
Why does the much older, less core, less threads, i7 outperform that later and bigger better R7?
Where do you see that? I don't even see it being tested. It's a "quad core" test specifically designed to not show performance, but to make for a useless report. Check overall speed, not a false "if we cripple it and don't actually test it" speed.
So from that test, it sure looks like the AMD is blowing the Intel away. Not just from the numbers you showed (those numbers are produced from 33% of the Intel, but only 25% of the AMD), but also from the fact that people are making these kinds of test just to make false numbers to try to make the Intel look competitive!
That's how slow the Intel is, that this kind of test even exists.
Oh so fake tests.
Yeah. It's a test engineered to promote one proc type over the other. Intel has long tuned their procs for legacy workloads because AMD is so far in front of them for general workloads. Intel has always been better at clock speeds and struggled with architecture, so they leverage that and it is sensible. Intel and AMD focus on different things.
Intel plays heavily towards video game and legacy "single or limited thread" workloads (video games rarely use more than a few, and you don't multi-task while gaming.) These are the big money spenders, anyway.. games and legacy users know that their computers are going to cost way more than normal because of their workloads.
AMD focuses on general case normal office use and normal server use and shines there. That's why in the Forbes test, the Intel was better for the video games (normally) and AMD for the business stuff. And that's still without balancing for actual use. Tests struggle to show multi-tasking performance even when done well. Once the cost was factored, though, AMD would have blown Intel away for every single test, not just the ones where it is tuned for it.
AMD is all about the office worker, the Linux server, and those kinds of things. Intel also went hard into collaboration with Microsoft to push Windows licensing to favour fewer, faster cores to more, slower cores. So while if you test server performance in pure hardware, AMD has long been the favourite. But if you factor in Windows licensing, Intel makes more sense. But if you move to Linux, you go back to AMD.
-
@Pete-S that's an interesting breakdown of how they price them and where the models compare over time.
-
Looking at the test you posted it says that in general benchmarks the Intel 8700K is 8% faster.
Lets just say that that is an indisputable fact without making any assumptions on what the tests actually test.Now it says that the AMD cpu is priced at kr 2594 and the Intel at kr 3311. That's means that the Intel is 28% more expensive.
So you pay 28% more for 8% higher performance - that means that the AMD 2700X is a better pick than the Core i7-8700K.
-
@Pete-S said in AMD Vs Intel:
Looking at the test you posted it says that in general benchmarks the Intel 8700K is 8% faster.
Lets just say that that is an indisputable fact without making any assumptions on what the tests actually test.Now is says that the AMD cpu is priced at kr 2594 and the Intel at kr 3311. That's means that the Intel is 28% more expensive.
So you pay 28% more for 8% higher performance - that means that the AMD 2700X is a better pick than the Core i7-8700K.
Looking at the prices today, it was a lot more than 28% more expensive. Closer to 40%.
-
@scottalanmiller said in AMD Vs Intel:
@Pete-S said in AMD Vs Intel:
Looking at the test you posted it says that in general benchmarks the Intel 8700K is 8% faster.
Lets just say that that is an indisputable fact without making any assumptions on what the tests actually test.Now is says that the AMD cpu is priced at kr 2594 and the Intel at kr 3311. That's means that the Intel is 28% more expensive.
So you pay 28% more for 8% higher performance - that means that the AMD 2700X is a better pick than the Core i7-8700K.
Looking at the prices today, it was a lot more than 28% more expensive. Closer to 40%.
I was going by the screenshot from the test.
Price might vary from day to day and market to market.
I remember a long time ago we bought a couple of HPE servers with E3 Xeons in them and they were so cheap that it was like you paid market price for the CPU and got the rest of the server for free. So if you buy machines (and not CPUs) you should compare the price of the machines against each other, and not the CPU per se. Sometimes you can get a good deal that will skew the results toward an Intel or AMD.