error: 3900 unable to allocate real <tmp>

[Jeong-Kweon Park]

Hello,

I am using Stata/MP 15 on Windows 10.
I have a question about my code.
I had the following error while running my code:

*: 3900 unable to allocate real <tmp>[536819,536819]
<istmt>: - function returned error

Perhaps I think it is due to 'matsize'.
Please let me know if you know how to solve this problem.

use "c:\temp\bmi.dta", clear

keep person_id last_date endpoint_hcc first_date bmi ///
         sex  age_cat  smk2 alccat_miss  ht_code hc_code dm_code cirrhosis
//==============================================================================
// #1: Set survival data for disease free survival analysis

stset last_date , failure(endpoint_hcc ==1) id(person_id) origin(first_date) scale(365.25)

//==============================================================================
// #2: Cox models participants and generate a graph (albumin)

*create spline variables for depression score
mkspline spline = bmi, nknots(4) displayknots cubic 

* select 10th percentile as the reference 
sum bmi, d
gen temp_p1      = r(p1)
gen temp_p10  = r(p10)
gen temp_p99  = r(p99)

sort bmi
gen bmi_ref50 = round(_N*0.5,0)
local ref50 bmi_ref50
gen temp_ref50 = bmi[`ref50']

foreach var of varlist spline* {
        gen temp_`var'_ref50 = `var' - `var'[`ref50']
        }

stcox spline* ib2.sex  i.age_cat  ib1.smk2 i.alccat_miss  ht_code hc_code dm_code cirrhosis

matrix temp_beta=e(b)
matrix beta=temp_beta[1,1..3]
matrix list beta
svmat beta, names(temp_beta)

matrix temp_var=e(V)
matrix var=temp_var[1..3,1..3]
matrix list var
svmat var, names(temp_var)

mata: xt_mat = st_data(., "temp_spline1_ref50 temp_spline2_ref50 temp_spline3_ref50")
mata: beta_mat = st_data(1, "temp_beta1 temp_beta2 temp_beta3")
mata: var_mat = st_data(1::3, "temp_var1 temp_var2 temp_var3")

mata: y   = xt_mat * beta_mat'
mata: se  = sqrt(diagonal(xt_mat * var_mat * xt_mat'))
    
mata: st_addvar("double", ("y","se"))
mata: st_store(., "y", y)
mata: st_store(., "se", se)
    
gen low = y - 1.96 * se
gen high  = y + 1.96 * se    

keep bmi y low high temp_ref50
saveold "spline.dta", version(12) replace

https://s3.ap-northeast-2.amazonaws....180510/bmi.dta