Userbenchmark new tantrum on upcoming Ryzen 7000

[SpyderTracks]

So based on some early benchmark performance numbers of the 7600x, Userbenchmark have seriously thrown in the towel. Looks like they'll be doing what they did with the 3000 series Ryzen and changing their scoring to make Intel more relevant "because Ryzen is too fast".

It's so rediculous you couldn't make it up

Source: https://cpu.userbenchmark.com/SpeedTest/1898605/AMD-Eng-Sample--100-000000593-20-Y

This is all based on an actual Userbenchark install and bench taken on a 7600x golden sample chip recently where the scores were pitted against the 5600x. Bear in mind, announcement due this week means that chips will be in production with first batches likely ready, which means this golden sample is more than likely the production performance.

UserBenchmark: AMD Advanced Marketing Devices 7600X vs Ryzen 5 5600X

 

[SpyderTracks]

And the related stories. Bear in mind, still no concrete announcements or verified 3rd party tests, userbenchmark is NOT a valid benchmark in any respect and this should not be considered any kind of metric of performance whatsoever.

All this does is highlight how rediculous userbenchmark are in their validity of testing, and obviously a paid intel shill not interesting in impartial analysis.

Ryzen 5 7600X Beats i9-12900K by 22% in New Single-Core Benchmarks

Zen 4 is looking dangerous

www.tomshardware.com

UserBenchmark goes to war with AMD as Ryzen 5 7600X takes down Intel Core i9-13900 in single-core speed comparison

The recent appearance of an AMD Ryzen 5 7600X engineering sample (ES) on UserBenchmark has sparked a series of events that has left the controversial synthetic benchmark taking cheap potshots at Team Red. The Ryzen 5 7600X even managed to outscore a Raptor Lake i9-13900 ES, leading UserBenchmark...

www.notebookcheck.net

AMD Ryzen 5 7600X 6-core Raphael CPU is up to 22% faster than Core i9-12900K in leaked single-core test - VideoCardz.com

AMD Ryzen 5 7600X 6-core/12-thread Raphael CPU An engineering sample with an AMD OPN code of 100-000000593-20_Y is now making rounds on Userbenchmark website This new processor has been tested with ASRock N7-B65XT motherboard, which is most likely a subvariant of the B650E motherboard designed...

videocardz.com

 

[Deleted member 41971]

because Ryzen is too fast".

not seeing the problem

 

[SpyderTracks]

I predicted that Ryzen 7000 would be roughly 15% increased performance over Intel 13th Gen (their upcoming new chips)

If this 56% improvement over Ryzen 5000 is accurate, there may be a hint of validity in that. I can't see Intel gaining the necessary 40% uplift on this generation.

Remember here they're comparing the 7600x to the i9 12900k, but the direct tier to tier competition would be the 7900x against the 12900k. 7600x is on par with a Core i5 like the 12600k

 

[Deleted member 41971]

I predicted that Ryzen 7000 would be roughly 15% increased performance over Intel 13th Gen (their upcoming new chips)

If this 56% improvement over Ryzen 5000 is accurate, there may be a hint of validity in that. I can't see Intel gaining the necessary 40% uplift on this generation.

Remember here they're comparing the 7600x to the i9 12900k, but the direct tier to tier competition would be the 7900x against the 12900k. 7600x is on par with a Core i5 like the 12600k

what im looking forward to is new amd, new nvidia 4000, new ddr5 ram etc etc, should be a sweet setup.

 

[SpyderTracks]

what im looking forward to is new amd, new nvidia 4000, new ddr5 ram etc etc, should be a sweet setup.

Yep, me too, this is my next rig, definitely although I may well go AMD on the GPU if rumours turn out to be true.

 

[Deleted member 41971]

Yep, me too, this is my next rig, definitely although I may well go AMD on the GPU if rumours turn out to be true.

how is amd gpu;s these days? I ask as am I right in saying amd had driver problems playing games and this meant nvidia was better in that respect.

 

[SpyderTracks]

how is amd gpu;s these days? I ask as am I right in saying amd had driver problems playing games and this meant nvidia was better in that respect.

That was with the 5000 series GPU's, had major driver issues, they never really got it right, think there was something inherently wrong with the architecture.

6000 series appear to have been far better. Lisa Su realised that driver stability was a major area that needed addressing so invested heavily in getting really high tier software developers a few years ago.

From what I've heard it's really paid off.

Especially with the new 2.0 version of FidelityFX Super Resolution which remember currently is purely software based in the driver, but is now offering actual competition with NVidia with regards to Ray Tracing and DLSS type performance. But the rumours are that the 7000 series GPU's will have hardware level RT cores much like NVidia's cards which should significantly boost Ray Tracing performance even further. But the main draw for me is rumours of equivalent performance to NVidia 4000 series using half the power.

Massive AMD Radeon RX 7000 series information dump reveals potential SKUs, hardware specifications, and launch date

Serial leaker Paul from RedGamingTech is at it again with a plethora of information regarding the upcoming AMD RDNA 3 graphics cards. The information includes the complete SKU list including alleged hardware specifications. The leaker has also revealed the launch date of different RDNA 3 GPUs.

www.notebookcheck.net

AMD's FidelityFX Super Resolution 2.0 really looks better than native 4K in Deathloop

And it offers a chunk more frames per second.

www.pcgamer.com

AMD FSR (FidelityFX Super Resolution): everything you need to know | Digital Trends

FidelityFX Super Resolution is AMD's answer to DLSS. Here's everything you need to know about FSR and how it works.

www.digitaltrends.com

 

[Deleted member 41971]

But the main draw for me is rumours of equivalent performance to NVidia 4000 series using half the power.

yep that would be a big draw, I think nvidia are going a bit overboard with how much power they require, especially now with electric costs.

 

[SpyderTracks]

yep that would be a big draw, I think nvidia are going a bit overboard with how much power they require, especially now with electric costs.

Yeah, absolutely.

We'll see, but the things that were preventing me considering AMD were:

1/. Driver stability - appears to be far better as of this generation
2/. RT and DLSS performance - appears to be far more competitive, and that's at a pure software layer, not including upcoming hardware RT cores on next gen gpu's

But I guess it all comes down to pricing. If AMD somehow pull off performance ABOVE NVidia (which I don't think anyone is expecting), then they may well hugely inflate their prices as a result which I guess they have done in some respect with their desktop processor line.

 

[Scott]

I would love a whole team red build to be honest. Especially with considering the infinity fabric etc.

There are a number of considerations for me too.

1. DLSS & RT. This is a must nowadays. We need performance and pretty with the money we put into these builds.
2. Infinity Fabric - Lets see some gains! This has to go hand in hand in some way with the above too I imagine.
3. SmartAccess (Direct Storage). This has to be just as good as Direct Storage. It has to be.

I'm not so concerned about the drivers now.

 

[SpyderTracks]

The tantrum continues for the 9800X3D chips!

Yet again, simping or admitting they've done zero testing of their own, 1% lows on the 9800X3D tend to outperform a lot of other titles average frame rate scores. It's streets ahead of anything for 1% lows, that's aside from the GPU performance it unlocks even at higher resolutions.

 

[TonyCarter]

Simply doesn’t understand that everyone tests at 1080p to give the CPU as much workload as it can without the GPU becoming the bottleneck. If you tested at 4k with a 4060 then the GPU will become the bottleneck and they’d they’d probably assume all CPUs are useless.

it must be the biggest troll farm on the planet…and I keep reporting search results with it in to google search for ‘misinformation’…as their ‘benchmarks’ are no such thing, I think that their tool is just a wrapper around a database that calculates a score based on frequency and cores…as I’ve seen some older CPUs with higher clocks (but much lower IPC) given higher ratings.

 

[SpyderTracks]

as their ‘benchmarks’ are no such thing, I think that their tool is just a wrapper around a database that calculates a score based on frequency and cores…as I’ve seen some older CPUs with higher clocks (but much lower IPC) given higher ratings.

That's exactly it, before Ryzen 3000 they actually marked higher core counts down (as AMD had far higher core counts) and purely scored on frequency, then when Intel moved to big / little design, they changed the metric to favour intel.

It's literal garbage, total nonsense, all their "benchmark tool" does is gather system specifications and broadcast them back to the main database to collate scores per component.

The irony that Intel themselves banned UserBenchmark from their own official reddit some years ago!

https://www.reddit.com/r/intel/comments/g36a2a