VOGONS

Common searches


To end the AMD v. Intel debate.

Topic actions

  • This topic is locked. You cannot reply or edit posts.

Reply 100 of 181, by awgamer

User metadata
Rank Oldbie
Rank
Oldbie

For several generations AMD on paper should match or beat intel/nvidia counterparts but always come up short, like having way more bandwidth and units but nvidia still performing better, they need to recalibrate, lead the target and take their less optimized results into account. A lot of the youtube tech channels are claiming AMD is trouncing Intel with their latest even in games while their benchmarks show otherwise, Intel is still top dog. I don't think they're paid off, more like hyperbole to get views. From what news I've seen ryzen 4000 should get closer to parity but still falling short here and there, while the news about upcoming intel is that they're claiming a big ipc jump to keep AMD from taking the crown. I'm all for intel finally doing a non incremental speed increase but they need to fix/remove all their security bugs/deliberate back doors. Other things I'd like to see from AMD, fix their multi monitor power usage, they fixed it with the radeon vii then broke it again with 5700 series back to pulling 30-40watts(wtf,) upgrade their cpu memory connection so it isn't crippled to just DDR4 3200(again, wtf, only reasonable excuse is rush to market) or even better bring quad channel to mainstream desktop, not just server/workstation.

Reply 101 of 181, by Standard Def Steve

User metadata
Rank Oldbie
Rank
Oldbie

Actually Zen2 in the Ryzen 3000 series is slightly faster per clock than Coffee Lake. Unfortunately the AMD chips just can't clock quite a high, so overall Intel is still slightly ahead in gaming. That's about it though. You see the new Zen2 based Ryzens and Threadrippers ahead of Intel in nearly everything else.
What AMD was able to pull off going from Zen+ to Zen2 is nothing short of spectacular. Their true Sandy Bridge moment, IMHO. 😀

Standard Def Rigs
Super P3: PIII-S @ 1.63 GHz/FSB155 | 2GB DDR-310 | 6800GT AGP | 500GB 7200 RPM
Super G4: 2x PowerPC 7455 @ 1.5 GHz | 2GB DDR-333 | 7800GS AGP | 300GB 10k RPM
Super G5: 4x PowerPC 970 @ 2.5 GHz | 16GB DDR2-533 | x1950XT PCIe | 512GB SSD

Reply 102 of 181, by mothergoose729

User metadata
Rank Oldbie
Rank
Oldbie
Standard Def Steve wrote:

Actually Zen2 in the Ryzen 3000 series is slightly faster per clock than Coffee Lake. Unfortunately the AMD chips just can't clock quite a high, so overall Intel is still slightly ahead in gaming. That's about it though. You see the new Zen2 based Ryzens and Threadrippers ahead of Intel in nearly everything else.
What AMD was able to pull off going from Zen+ to Zen2 is nothing short of spectacular. Their true Sandy Bridge moment, IMHO. 😀

Zen+ was quite competitive on 12nm. I believe that Intel's 14nm is actually slightly smaller than TSMC 12nm, so intel had a node advantage. TSMC 7nm is actually larger than Intel's 10nm, but intel shit the bed on that one, so now AMD has a good architecture and a node advantage. Good on AMD for being in a position to capitalize on their advantage. They certainly wouldn't be ahead if they were still on a bulldozer/pile driver architecture.

I do look forward to what Intel will deliver when it finally delivers a new architecture on a new processing node. I personally use my PC most for emulation, and emulation loves high per thread performance. According to leaked slides Intel's next architecture will offer as much as 40% more IPC in some workloads and 18% better IPC overall, which would be the biggest jump we have had seen since... sandy bridge.

Reply 103 of 181, by Scali

User metadata
Rank l33t
Rank
l33t
Standard Def Steve wrote:

Actually Zen2 in the Ryzen 3000 series is slightly faster per clock than Coffee Lake. Unfortunately the AMD chips just can't clock quite a high, so overall Intel is still slightly ahead in gaming.

I'm not so sure if clockspeed is the answer here.
IPC is not a constant, it depends very much on the exact job at hand.
The instruction mix makes a difference. Also, there's more to game performance than just the CPU. There's also the GPU, which is connected through the PCI-e bus.
So I wouldn't rule out the possibility that Intel's advantage in gaming is not so much the result of raw IPC or raw clockspeed, but also perhaps the whole I/O interface to the GPU.
Because if I see comparisons between Zen2 and Core at the same base speeds (3.6 GHz), the Core has a slight advantage. It may be able to turbo a few hundred MHz higher, but would it actually do that during a game benchmark? I would more or less expect the game to run the CPU so hot that it won't be able to sustain turbo speeds anyway. So I wonder if there actually is a clock speed advantage in these benchmarks.

http://scalibq.wordpress.com/just-keeping-it- … ro-programming/

Reply 104 of 181, by awgamer

User metadata
Rank Oldbie
Rank
Oldbie

I'd be interested in seeing some comparisons(clock for clock and max) with dosbox and mame of coffeelake-s(9900k/ks) and ryzen 3k(3700x-3950x.) Oh yeah, AMD also needs to fix their clock speeds never hitting rated speed for any reasonable duration in order to claim that speed(a second or two intermittently doesn't count AMD.)

Reply 105 of 181, by mothergoose729

User metadata
Rank Oldbie
Rank
Oldbie
Scali wrote:
I'm not so sure if clockspeed is the answer here. IPC is not a constant, it depends very much on the exact job at hand. The inst […]
Show full quote
Standard Def Steve wrote:

Actually Zen2 in the Ryzen 3000 series is slightly faster per clock than Coffee Lake. Unfortunately the AMD chips just can't clock quite a high, so overall Intel is still slightly ahead in gaming.

I'm not so sure if clockspeed is the answer here.
IPC is not a constant, it depends very much on the exact job at hand.
The instruction mix makes a difference. Also, there's more to game performance than just the CPU. There's also the GPU, which is connected through the PCI-e bus.
So I wouldn't rule out the possibility that Intel's advantage in gaming is not so much the result of raw IPC or raw clockspeed, but also perhaps the whole I/O interface to the GPU.
Because if I see comparisons between Zen2 and Core at the same base speeds (3.6 GHz), the Core has a slight advantage. It may be able to turbo a few hundred MHz higher, but would it actually do that during a game benchmark? I would more or less expect the game to run the CPU so hot that it won't be able to sustain turbo speeds anyway. So I wonder if there actually is a clock speed advantage in these benchmarks.

Intel's monolothic architecture has some advantages when it comes to intercore bandwidth and cache latency, but from what I have seen in benchmarks Intel's processors are clocked 10-15% faster than AMD, and they perform as much as 10-15% in games. If there is an archtectural advantage for gaming that Intel enjoys, it seems ever so slight. It has been said before, overall AMD enjoys a very small IPC lead at the moment, in particular in whatever makes blender and Cinebench go. They consistently do really well in those benchmarks.

Reply 106 of 181, by awgamer

User metadata
Rank Oldbie
Rank
Oldbie

Maxed out Intel vs maxed out AMD, Intel has a 15-30% advantage in enough games for that to be the metric to beat. When nvidia releases their next GPU soon(AMD is calling their next one the nvidia killer but then they've said that before and failed to deliver,) will remove the gpu as bottleneck more making the CPU differential stand out further like when they show lower res benchmarks (https://www.guru3d.com/articles-pages/intel-c … -review,20.html though this is only showing stock, not max/overclocked,) it will be interesting to see how well the upcoming improved AMD and Intel CPUs fare in that environment.

Reply 107 of 181, by SPBHM

User metadata
Rank Oldbie
Rank
Oldbie
Firtasik wrote:
DNSDies wrote:

Firtasik's charts had a $1100 Intel CPU topping the charts of a few games with Ryzen 3800x about 7-8FPS behind it, which costs about $300.

i5 9600K costs about $200 and it beats 3800x (about $370) in some games.

if you are going that route you could say the 3600 costs $200 and it beats the 9600K in some games and in pretty much all other applications

in any case almost all of my PCs are using Intel CPUs, because they were the best deal at the time, right now I would have a hard time not going AMD for new parts.

Reply 109 of 181, by SPBHM

User metadata
Rank Oldbie
Rank
Oldbie
Firtasik wrote:

i5 9600K is generally faster than R5 3600 in gaming:
https://www.gamersnexus.net/hwreviews/3527-am … nchmark-in-2019

it is, but not in every game, and not to a huge extent in most cases, and as time goes by the advantage should swing more towards the 3600 due to SMT being a pretty big advantage for more heavily multi threaded applications, just check the first graph of the test;

and AM4 has better upgrade options

Reply 111 of 181, by Scali

User metadata
Rank l33t
Rank
l33t

Anyway, it's clear that Intel has some new products that have been delayed, so they have been 'stuck in the pipeline' for a while (not just CPUs, but they are also getting behind on the entire platform, with new PCIe standards and all).
I'd like to see what comes out once the pipeline is unclogged.
Intel should be coming out with the stopgap Comet Lake very soon now, which will be on their "14++ nm" node. It could be interesting... because Zen2 and Coffee Lake are so closely matched right now. Comet Lake could put Intel back in the lead. So I'd want to wait for that before I make my purchase.

As for me personally, I am kinda sorta, but not really in the market for a new PC. Intel still thinks they will release 10nm desktop parts in 2020. So I personally will wait for that as well. If Intel doesn't pull ahead on 10nm, then AMD will likely retain its advantage for a few years to come.

http://scalibq.wordpress.com/just-keeping-it- … ro-programming/

Reply 112 of 181, by Bruninho

User metadata
Rank Oldbie
Rank
Oldbie

Wait, so in the end I didn’t needed an i7 for serious gaming, just an overclocked i5 for a cheaper price? *shocked*

"Design isn't just what it looks like and feels like. Design is how it works."
JOBS, Steve.

List of ALL Android vulnerabilities

Reply 113 of 181, by brostenen

User metadata
Rank l33t++
Rank
l33t++

My first PC had an Cyrix 486slc2-50 and what a pile of shit that was. Sure I would like one again, yet that would only be because of the personal nostalgic attachment that I have to it. Yes.... First computer... Nostalgic trip.... Would last for like 1 day and then I am back on Amiga, C64 and Amd/Intel 486's.

I don't really care if it is Amd or Intel. At the right frequency, they both run the same software. Just give me the cheapest version of the CPU and I am a happy champ.

Don't eat stuff off a 15 year old never cleaned cpu cooler.
Those cakes make you sick....

My blog: http://to9xct.blogspot.dk
My YouTube: https://www.youtube.com/user/brostenen

001100 010010 011110 100001 101101 110011

Reply 114 of 181, by mothergoose729

User metadata
Rank Oldbie
Rank
Oldbie
Scali wrote:
Anyway, it's clear that Intel has some new products that have been delayed, so they have been 'stuck in the pipeline' for a whil […]
Show full quote

Anyway, it's clear that Intel has some new products that have been delayed, so they have been 'stuck in the pipeline' for a while (not just CPUs, but they are also getting behind on the entire platform, with new PCIe standards and all).
I'd like to see what comes out once the pipeline is unclogged.
Intel should be coming out with the stopgap Comet Lake very soon now, which will be on their "14++ nm" node. It could be interesting... because Zen2 and Coffee Lake are so closely matched right now. Comet Lake could put Intel back in the lead. So I'd want to wait for that before I make my purchase.

As for me personally, I am kinda sorta, but not really in the market for a new PC. Intel still thinks they will release 10nm desktop parts in 2020. So I personally will wait for that as well. If Intel doesn't pull ahead on 10nm, then AMD will likely retain its advantage for a few years to come.

The rumor is that Intel is abandoning their 10nm node entirely for desktop and HEDT. They won't release high performance parts until 2022 - on a 7nm node

https://www.hardwareluxx.de/index.php/news/ha … p-komplett.html

Zen 3 is set to release next year, and is rumored to have 8% higher IPC and a 200mhz higher clock speed than Zen 2. According to their roadmap, Zen 4 is slated for 2022, and if the rumors about Intel's 10nm node are true, this is the architecture intel will be competing with when they roll out a 7nm node of their own.
https://wccftech.com/amd-zen-3-to-deliver-8-i … umors-indicate/

Meanwhile, TSMC is making progress on a 5nm node. Optimistically, could be available as soon as 2021. If so, Zen 4 will likely be a 5nm processor.

https://www.pcgamesn.com/amd/tsmc-5nm-zen-4

Reply 115 of 181, by appiah4

User metadata
Rank l33t++
Rank
l33t++
Scali wrote:
Anyway, it's clear that Intel has some new products that have been delayed, so they have been 'stuck in the pipeline' for a whil […]
Show full quote

Anyway, it's clear that Intel has some new products that have been delayed, so they have been 'stuck in the pipeline' for a while (not just CPUs, but they are also getting behind on the entire platform, with new PCIe standards and all).
I'd like to see what comes out once the pipeline is unclogged.
Intel should be coming out with the stopgap Comet Lake very soon now, which will be on their "14++ nm" node. It could be interesting... because Zen2 and Coffee Lake are so closely matched right now. Comet Lake could put Intel back in the lead. So I'd want to wait for that before I make my purchase.

As for me personally, I am kinda sorta, but not really in the market for a new PC. Intel still thinks they will release 10nm desktop parts in 2020. So I personally will wait for that as well. If Intel doesn't pull ahead on 10nm, then AMD will likely retain its advantage for a few years to come.

Intel and AMD are closely matched only in your strange fantasyland. And there is not a chance in hell Cometlake will be anything but a şacklustre refresh.

Retronautics: A digital gallery of my retro computers, hardware and projects.

Reply 116 of 181, by appiah4

User metadata
Rank l33t++
Rank
l33t++
Firtasik wrote:

i5 9600K is generally faster than R5 3600 in gaming:
https://www.gamersnexus.net/hwreviews/3527-am … nchmark-in-2019

Yes but it is shit for anything else but gaming.

Ill let this be a turnousol test for detecting idiots. What would you rather have? The cpu that has a fleeting marginal advantage in gaming but sucks balls in everything else or a CPU that is all around amazingly good at everything?

Retronautics: A digital gallery of my retro computers, hardware and projects.

Reply 117 of 181, by Scali

User metadata
Rank l33t
Rank
l33t
appiah4 wrote:

Intel and AMD are closely matched only in your strange fantasyland.

Pretty sure the context of the discussion was gaming, and pretty sure the benchmarks support that they are closely matched in gaming.
If the rumours of Comet Lake-S being 8-18% faster than Coffee Lake are true, then that would give Intel a lead of about 8-18% (while also making them less 'suck balls' in everything else). Since these CPUs can be introduced any day now, I think it's worth it to wait and see if the rumours are true.

Last edited by Scali on 2019-12-08, 14:22. Edited 3 times in total.

http://scalibq.wordpress.com/just-keeping-it- … ro-programming/

Reply 118 of 181, by Scali

User metadata
Rank l33t
Rank
l33t
mothergoose729 wrote:

The rumor is that Intel is abandoning their 10nm node entirely for desktop and HEDT. They won't release high performance parts until 2022 - on a 7nm node

Depends on which rumours you want to include/exclude I suppose 😀
This was a rumour/leak 3 days ago, about the 14++ nm Comet Lake-S I mentioned: https://wccftech.com/intel-comet-lake-desktop … k-z490-spotted/

And in that same article they also link to this article from Nov 1st: https://wccftech.com/intel-10nm-desktop-cpus- … ing-early-2020/
And it claims to include an official statement from Intel, which is allegedly a response to the rumour that Intel would have abandoned 10 nm altogether, which very specifically confirms desktop SKUs on 10 nm next year.
So 2020 or 2022? You tell me.

http://scalibq.wordpress.com/just-keeping-it- … ro-programming/

Reply 119 of 181, by appiah4

User metadata
Rank l33t++
Rank
l33t++
Scali wrote:
appiah4 wrote:

Intel and AMD are closely matched only in your strange fantasyland.

Pretty sure the context of the discussion was gaming, and pretty sure the benchmarks support that they are closely matched in gaming.
If the rumours of Comet Lake-S being 8-18% faster than Coffee Lake are true, then that would give Intel a lead of about 8-18% (while also making them less 'suck balls' in everything else). Since these CPUs can be introduced any day now, I think it's worth it to wait and see if the rumours are true.

Considering parts they marketed as 5ghz turned out to be 3.8ghz parts I consider anything that comes out of Intels mouth to be a lie.

Retronautics: A digital gallery of my retro computers, hardware and projects.