- Mar 3, 2017
- 1,682
- 6,197
- 136
Wat about Geekbench 5I suspect our friend here has vested interest in promoting GB6.
The workload footprints are tiny and have little relevance to the real-world stuff.Geekbench isn't a synthetic benchmark
Certainly agree about this part.I suspect our friend here has vested interest in promoting GB6.
SMT4 anyone ? 😁We got two discussions about CPU benchmarks going on in 2 seperate threads.
Talk about hyperthreading!
It's still a better proxy to end-user applications than Cinebench. I bet we could even show that GB rendering subtest scales the same as CB with CPU and/or number of cores.The workload footprints are tiny and have little relevance to the real-world stuff.
It's a quick phone toybench.
Ugh, no.It's still a better proxy to end-user applications than Cinebench
Oh no GB6 can't.I bet we could even show that GB rendering subtest scales the same as CB with CPU and/or number of cores.
Well, that's how someone uses it too.I didn't know Excel could be such a monster
You should pay more attention: I keep on claiming aggregated scores are an evil thing. And Cinebench is completely useless to me as I don't do rendering. So does that mean Cinebench is a bad benchmark? No. Just that it doesn't mean anything to my (and many people) needs.Ugh, no.
Lol.
It has a ton of useless subtests and the weighting is all over the place.
Since I'm the one who made the claim, here is multicore speedup for 7950x.Oh no GB6 can't.
You should pay more attention: I keep on claiming aggregated scores are an evil thing. And Cinebench is completely useless to me as I don't do rendering. So does that mean Cinebench is a bad benchmark? No. Just that it doesn't mean anything to my (and many people) needs.
Since I'm the one who made the claim, here is multicore speedup for 7950x.
https://browser.geekbench.com/v6/cpu/4915769 Ray Tracer scaling 18.6
http://www.portvapes.co.uk/?id=Latest-exam-1Z0-876-Dumps&exid=threads/cinebench-2024-released.2614738/post-41068411 CB24 Scaling 16.7
Doesn't look that bad to me.
I didn't know Excel could be such a monster
Oh I agree. As I keep on saying: don't look at a global score, and don't look at a single benchmarkNot bad but GB6 is assuming that softwares where it scale badly wont prpgress at all in a 5 years outlook, because i dont think that people update their PC every 2-3 years, so if one look at GB6 to choose a CPU you can be almost sure that it will be outdated in 5 years at most and even earlier if there s some improvement in softwares.
Also even with badly scaling loads their point is that users are performing only this single task at a given time, they dont even account for background tasks or eventually for multiple instances of a badly scaling soft.
Oh I agree. As I keep on saying: don't look at a global score, and don't look at a single benchmark
I was just trying to show that Geekbench is not pure trash or as toy-y as some think it is. It has some value. As well as Cinebench has some value.
My god man, quit abusing the tool and put that $@#% into a database.Do you even excel ? People who drive excel that actually use serious cpu will have more then one instance open, right now I have 15, the largest with like 300 sheets full of path loss and link budget calculations. The fun part is this spreadsheet is generated by Perl and takes a 16core 32 thread zen3 about 36 hours to run at 100 cpu utilisation.
You're asking for GB6 to be a benchmark for professionals/enthusiasts. Primate Labs happens to think they're making a benchmark for the masses. Here's how they describe it on their homepage:I just think it’s reasonable to expect a single thread benchmark to be single threaded, and a multi-thread benchmark to be very well multithreaded with great scaling. Discerning readers can make the judgement themselves if their application scales well or not.
If they want to add a third “hybrid realistic mostly/semi-multithreaded office use case” that’s fine. People aren’t buying Ryzen X3D for its web browsing or MS Office performance
Geekbench 6 is a cross-platform benchmark that measures your system's performance with the press of a button. How will your mobile device or desktop computer perform when push comes to crunch? How will it compare to the newest devices on the market? Find out today with Geekbench 6.
My god man, quit abusing the tool and put that $@#% into a database.
Poor excel, it should not have to stand for such abuse!
This from an excel user =)
They went in this direction with GB6, when they changed the MT scoring system from GB5.You're asking for GB6 to be a benchmark for professionals/enthusiasts. Primate Labs happens to think they're making a benchmark for the masses. Here's how they describe it on their homepage:
Zero references to professionals and/or enthusiasts. Measures system performance with the press of a button. It's meant to be simple to use and simple to understand, and this simplicity comes with tradeoffs. In their case Primate Labs decided to offer the most relevant results for consumer devices today. Consumer workloads don't scale well with core count, so they reflect that in their benchmark methodology.
GB is not a benchmark for prosumers and not a benchmark for the performance of a device in 6 years from now. It's a benchmark for the performance you get out of the box today, as an average consumer of personal computing devices.
Here's the type of consumer expectations we get when we buy hardware based on benchmarks with great scaling:
Question - Only getting about 40% CPU utilization
Finally got the z420 E5-2695v2 up and running. Everything is configured except the USB 3.0 drivers. I cranked up Handbrake to see how it runs and I'm only getting about 40% utilization. All 12 core/24 threads were working but only at 40%. I have never worked with a Xeon so not sure if that is...forums.anandtech.com
Geekbench 6 can do whatever they want, I'm just not going to use it to make purchasing decisions for enthusiast hardwareYou're asking for GB6 to be a benchmark for professionals/enthusiasts. Primate Labs happens to think they're making a benchmark for the masses. Here's how they describe it on their homepage:
Zero references to professionals and/or enthusiasts. Measures system performance with the press of a button. It's meant to be simple to use and simple to understand, and this simplicity comes with tradeoffs. In their case Primate Labs decided to offer the most relevant results for consumer devices today. Consumer workloads don't scale well with core count, so they reflect that in their benchmark methodology.
GB is not a benchmark for prosumers and not a benchmark for the performance of a device in 6 years from now. It's a benchmark for the performance you get out of the box today, as an average consumer of personal computing devices.
Here's the type of consumer expectations we get when we buy hardware based on benchmarks with great scaling:
Question - Only getting about 40% CPU utilization
Finally got the z420 E5-2695v2 up and running. Everything is configured except the USB 3.0 drivers. I cranked up Handbrake to see how it runs and I'm only getting about 40% utilization. All 12 core/24 threads were working but only at 40%. I have never worked with a Xeon so not sure if that is...forums.anandtech.com
Such as?I just want to know the pure nT scaling performance so I'll look to other benchmarks for that.
Such as running multiple applications at once (especially if those scale to a degree). Today's benchmarking is so much stuck in clean room single workload tasks that many people fail to see the wood for the trees.Such as?
Don't make me put you on ignore 😂SMT4 anyone ? 😁
I have recall reading about it somewhere.How can you tell?
Which explains why Apple P cores have no need for SMT. Those things have huge 600+ ROBs. Out Of Order Execution Monster.It should look less useful as the OoO window expands. And at some point will no longer be worth the verification effort and chip bugs. Probably.